• Open Access

Orthopedic bioactive implants: Hydrogel enrichment of macroporous titanium for the delivery of mesenchymal stem cells and strontium



Insufficient implant stability is an important determinant in the failure of cementless prostheses. To improve osseointegration, we aim at generating a bioactive implant combining a macroporous titanium (TT) with a biocompatible hydrogel to encapsulate osteo-inductive factors and osteoprogenitor cells. Amidation and cross-linking degree of an amidated carboxymethylcellulose hydrogel (CMCA) were characterized by FT-IR spectrometry and mechanical testing. Bone marrow mesenchymal stem cells (BMSCs) from osteoarthritic patients were cultured on CMCA hydrogels, TT, and TT loaded with CMCA (TT + CMCA) with an optimized concentration of SrCl2 to evaluate cell viability and osteo-differentiation. Amidation and cross-linking degree were homogeneous among independent CMCA batches. SrCl2 at 5 μg/mL significantly improved BMSCs osteo-differentiation increasing calcified matrix (P < 0.01), type I collagen expression (P < 0.05) and alkaline phosphatase activity. TT + CMCA samples better retained cells into the TT mesh, significantly improving cell seeding efficiency with respect to TT (P < 0.05). BMSCs on TT + CMCA underwent a more efficient osteo-differentiation with higher alkaline phosphatase (P < 0.05) and calcium levels compared to cells on TT. Based on these in vitro results, we envision the association of TT with strontium-enriched CMCA and BMSCs as a promising strategy to generate bioactive implants promoting bone neoformation at the implant site. © 2013 Wiley Periodicals, Inc. J Biomed Mater Res Part A: 101A: 3396–3403, 2013.


Joint replacement is widely used to recover articular functionality in joints compromised by degenerative pathologies or severe traumas. The integration between bone and implant is crucial in cementless prostheses where bone ingrowth is required to grant fixation. The achievement of an optimal osseointegration can lead to a reduction in the risk of implant mobilization and, consequently, in the number of revision procedures,[1-3] which have major drawbacks such as patients' morbidity and relevant costs for National Health Systems. With this aim, cementless implant technology has evolved with the introduction of porous metallic materials to maximize bone ingrowth and bone-to-implant contact.[4-7] Among the different types of porous titanium, Trabecular Titanium™ (TT) is currently clinically used in acetabular cups and to fill cavitary and segmental bone defects. This macroporous titanium is characterized by multiple layers of hexagonal pores with a 640 μm diameter,[8] macroporosity that facilitates cell migration and bone deposition.[9] It has been shown that porous titanium implants can be enriched, either directly[10-12] or by association with biocompatible hydrogels,[13-16] with antibiotics,[14] VEGF,[15] and FGF-2.[16] To improve implant osseointegration, osteogenic factors can be used for implant enrichment. Strontium is a low-cost ion and is easier to handle compared to osteo-inductive proteins, such as BMP-2, being therefore an ideal candidate to enrich off-the-shelf materials for bone tissue engineering, as proposed for hydroxyapatite[17, 18] and, more recently, for titanium.[19, 20] Indeed, strontium ranelate is a drug used for the treatment of osteoporosis, thanks to its ability to increase bone neoformation and to reduce bone resorption.[21-23] Furthermore, strontium has a positive effect on osteogenic differentiation of bone marrow mesenchymal stem cells (BMSCs),[24-28] which are resident at the implant site and are characterized by an intrinsic osteogenic potential.[29-31] It has been demonstrated that the use of autologous bone marrow concentrate, obtained by an intra-operative approach, improves bone healing.[32] Thus, loading BMSCs into the implant may represent a promising approach to increase the population of osteogenic progenitor cells[33, 34] at the implant site, especially when aging and age-related diseases negatively affect bone deposition.[35]

We envision that autologous BMSCs can be isolated and seeded into a titanium prosthesis preloaded with a strontium-enriched hydrogel with a completely intra-operative approach. The implantation of a bioactive prosthesis, delivering cells and osteogenic signals, could result in higher implant stability and faster osseointegration, reducing recovery time and improving patients' quality of life. In this study we combined TT with an amidated carboxymethylcellulose (CMCA) hydrogel, previously used for chondrocyte culture.[36] We evaluated if cell seeding efficiency and osteogenic differentiation were improved when TT was enriched with CMCA (TT + CMCA) as compared to normal TT that represents a clinical standard reference. To better resemble a possible clinical application, BMSCs were obtained from osteoarthritic (OA) patients.


Substrates preparation

CMCA polymer synthesis

A water-based synthesis was developed by the chemical laboratories of Limacorporate s.p.a. to obtain CMCA from carboxymethylcellulose (CMC, CEKOL® 30000, CP Kelko), in alternative to organic synthesis.[37] CMC was solubilized in ddH2O (1.5% W/V) with N-hydroxysuccinimide (Fluka) at a molar ratio 1 : 1. Methylamine (Merk) was added to CMC solution, at a molar ratio 1 : 1 with CMC. pH was adjusted to 4.7 with HCl (Sigma-Aldrich) and finally, amidation of CMC was obtained by adding 1-ethyl-3-(3-dimethyl-aminopropyl)-carbodiimide (EDC, Ubichem) at a molar ratio 1 : 1 with CMC. After 3 h at r.t., purification was performed by ultrafiltration (Millipore Cogent M1, 100 kDa membrane). Purity was controlled by FT-IR spectrometry (Spectrum 400 FT-IR FT-NIR, Perkin-Elmer). Purified CMCA was freeze-dried for storage.

Validation of CMCA polymer functionalization by FT-IR spectrometry

FT-IR spectrometry was used for a semiquantitative evaluation of amidation degree. As bands for COO (1590 cm−1) and amidic group (1640 cm−1) were too near to allow a clear distinction between them, an acidification step with Dowex resin (Sigma-Aldrich) was performed to convert COO groups into COOH groups. As a result, the absorption peaks of amidic groups (1640 cm−1) were clearly distinguished from those of the groups that did not react (1730 cm−1) and the percentage of amidic moieties introduced was calculated.

CMCA hydrogel cross-linking

Freeze-dried CMCA polymer was solubilized (1.7% W/V) in ddH2O together with N-hydroxysuccinimide at molar ratio 1 : 1 with CMCA. 1,3-diaminopropane (Merk) was added in a ratio of 1 : 2 with CMCA and pH was adjusted to 4.7 with HCl. Finally, the activator EDC was added at a molar ratio 1 : 1 with CMCA. Cross-linking reaction took 3 h and hydrogel purification was performed by washing in ddH2O. To verify cross-linking reproducibility, 5 min after the addition of EDC 6 mL of reaction blend were sampled with a 10 mL syringe without needle and kept at r.t. to let the cross-linking reaction end. As cross-linking degree affects hydrogel mechanical properties, LFPlus digital testing tensile machine (Amestek) was used to measure the injection force needed to extrude the hydrogel from the syringe.

CMCA hydrogel loading into TT mesh

TT (Ti6Al4V, Ø 8 mm, h 3 mm, pore Ø 640 μm, manufactured by Limacorporate s.p.a.) was obtained by electron beam melting.[8] Hydrated CMCA was injected into TT mesh through a syringe. The amount of hydrated CMCA loaded was determined by weighing (analytical balance Ohaus, Adventurer Pro AV2101C). Constructs were oven-dried at 50°C and sterilized by EtO.

BMSCs culture and osteogenic differentiation

BMSCs isolation and expansion

Bone marrow was harvested from the femoral compartment of 9 OA patients (mean age 59±4 years) undergoing total hip replacement, after written consent. Bone marrow was centrifuged and plated in control medium consisting of HG-DMEM (High Glucose, Gibco) with 10% Fetal Bovine Serum (FBS, Lonza), 0.029 mg/mL L-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 10 mM hepes, 1 mM sodium pyruvate (all from Gibco) supplemented with 5 ng/mL FGF-2 (Peprotech). BMSCs were selected by plastic adherence and expanded. Cells were frozen at passage 2 and, at need, thawed and expanded until passage 4.

Selection of SrCl2 concentration to improve BMSCs osteogenic differentiation

Cells were seeded at 3 × 103 cells/cm2 and cultured for 14 days in osteogenic medium consisting of control medium supplemented with 10 nM dexamethasone, 10 mM glycerol-2-phosphate, 150 μM L-ascorbic acid-2-phosphate and 10 nM cholecalciferol (all from Sigma-Aldrich)[38] in the presence of 0, 1, 5, 10, and 20 μg/mL SrCl2 (Sigma-Aldrich).

Culture of BMSCs on CMCA hydrogel

Hydrated CMCA was dispensed in a 24-multiplate (0.75 mL/well), and then oven-dried at 50°C and sterilized in EtO. Twenty-four hours before cell seeding, CMCA samples were hydrated with 0.75 mL of medium. A total of 7.5 × 105 BMSCs were suspended in 1 mL of medium, seeded on CMCA and cultured up to 21 days in different media: control, osteogenic, or osteogenic with 5 μg/mL SrCl2.

Osteogenic differentiation of BMSCs on TT and TT + CMCA

A total of 7.5 × 105 BMSCs were suspended in 75 μL of medium and seeded on normal TT or on TT + CMCA samples. After 3 h, 1 mL of medium was added. TT and TT + CMCA samples were cultured up to 21 days in osteogenic medium with 5 μg/mL SrCl2.

Evaluation of cell behavior on biomaterials

Cell seeding efficiency

Seeding efficiency on TT and TT + CMCA was determined by DNA quantification 24 h after seeding (CyQuant kit, Invitrogen) and calculated as percentage of the number of seeded cells.

Cell viability

Cell viability was determined by Alamar Blue assay (Invitrogen). 800 μL of 10% Alamar Blue in HG-DMEM w/o phenol red were added to each sample and incubated for 4 h at 37°C. Fluorescence (540–580 nm) was read using a Victor X3 Plate Reader (Perkin Elmer).

Calcified matrix deposition

Calcified matrix was quantified by Alizarin Red-S (AR-S, pH 4.1, Fluka) staining. After washing, each sample was unstained with 10% cetylpyridinium chloride monohydrate (CPC, Sigma-Aldrich) in 0.1M phosphate buffer (pH 7.0) and absorbance was read at 570 nm.[39]

Gene expression of type I collagen

Expression of type I collagen (COL1A1) was evaluated by real time PCR (Rotor Gene RG3000 system, Qiagen). RNA was purified with RNeasy Mini kit (Qiagen)and reverse-transcribed to cDNA with iScript cDNA Synthesis Kit (Bio-Rad Laboratories) (5 min at 25°C, 30 min at 42°C, 5 min at 85°C). 20 ng of cDNA were incubated with a PCR mixture including TaqMan Universal PCR Master Mix and TaqMan® Assays-on-Demand™ Gene expression probes (Life Technologies) (2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C, 1 min at 60°C). COL1A1 expression was normalized on glyceraldehyde 3-phosphate dehydrogenase (GAPDH).

Alkaline phosphatase activity

Alkaline phosphatase (ALP) activity was determined by enzymatic assay incubating cell lysates at 37°C with 1 mM p-nitrophenylphosphate in 100 mM diethanolamine and 0.5 mM MgCl2 (pH 10.5, Sigma-Aldrich).[40] Absorbance was read at 405 nm. ALP was normalized on protein content, determined by BCA Protein Assay Kit (Pierce Biotechnology).

Calcium quantification

Calcium quantification was performed by Randox assay (Randox Laboratories Ltd). Samples were incubated with 0.5M HCl for 6 h at 4°C on a rotating plate. Supernatants were incubated with the Randox working solution. Absorbance was read at 570 nm.

Scanning electron microscopy analysis

BMSCs adhesion and matrix production were evaluated by SEM. Samples were fixed for 1 h in glutaraldehyde (1.2% in 0.1M sodium cacodylate buffer), washed with 0.1M sodium cacodylate buffer and fixed for 1 h in OsO4 (1% in 0.1M sodium cacodylate buffer). Samples were dehydrated through graded EtOH, submitted to critical point drying with CO2, sputter coated with gold and analyzed with Sigma Scanning Electron Microscope (Zeiss, 10 kEV).

Statistical analysis

Data are expressed as mean ± SEM, unless differently specified. Statistical analyses were performed using Student's t test to compare two groups and Two-Way ANOVA to compare groups at different time points (Graph Pad Prism, v 5.0). Significance was set at P < 0.05.


Substrates preparation

Validation of CMCA hydrogel amidation and cross-linking

Hydrophilicity of CMC was increased by introducing amidic groups in the polymer backbone. To avoid the use of dimethylformamide, a very toxic reagent[41-43] used for the organic synthesis of CMCA,[37] a water-based synthesis was developed. FT-IR spectra [Fig. 1(a)] allowed the quantification of amidic groups introduced and, as shown in Figure 1b, we observed that the amidation degree was homogeneous among independent samples (39.4 ± 2.3%). Different CMCA batches were homogeneous for cross-linking degree as demonstrated by similar values of injection force (258.5 ± 12.8 N) measured by a mechanical extrusion test [Fig. 1(c)]. To demonstrate the reliability of this method, cross-linking was performed starting from different CMCA concentrations. When CMCA concentration was lower (1.6% W/V) injection force was reduced, whereas when it was higher (1.8% W/V) we observed an increase in this parameter. This method, easier and less expensive compared to titration and NMR,[37, 44] can be used as a routine quality control to evaluate the homogeneity of cross-linking among different batches of production and can reveal small differences in the starting conditions of the cross-linking reaction. Nevertheless, quality controls should include further analyses to grant the spatial homogeneity of hydrogel cross-linking.

Figure 1.

CMCA amidation and cross-linking degree. (a) FT-IR spectra of CMCA before and after the acidification step (CMCA_H). Inset represents COOH and amidic group peaks. (b) Amidation degree calculated on the basis of FT-IR spectra (mean ± SD, n = 12). (c) Cross-linking validation by mechanical testing starting from 1.7% CMCA (W/V, mean ± SD, n = 12). Control samples were prepared starting from 1.6% and 1.8% CMCA (W/V).

CMCA hydrogel loading into TT mesh

CMCA loaded within the TT mesh through a syringe was homogeneously distributed in TT pores [Fig. 2(a)]. CMCA loading was reproducible as demonstrated by the low variability in the amount of hydrogel loaded among different samples [Fig. 2(b)]. In our approach the interaction between CMCA and TT was granted by the increase of volume of CMCA that, once rehydrated with the cell suspension, remained entrapped within TT structure without any loss of hydrogel during culture. As an alternative, phosphonate derivatives of CMC can be developed to generate a chemical bound between titanium and hydrogel.[45]

Figure 2.

CMCA loading into TT. (a) Section of TT + CMCA (scale bar 5 mm, CMCA stained red). (b) Weight of hydrated CMCA loaded into TT mesh (mean ± SD, n = 47). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Selection of SrCl2 concentration to improve BMSCs osteogenic differentiation

Aiming at the future incorporation of SrCl2 within CMCA hydrogel, BMSCs were cultured in the presence of different amounts of SrCl2 to select a suitable concentration for the improvement of their osteogenic differentiation. As these in vitro tests are oriented towards our aim to implant a bioactive device made from TT and CMCA directly into the bone, all the experiments were performed in osteogenic medium, mimicking the differentiative signals that in vivo would be provided by the adjacent bone.

The evaluation of osteogenic markers in BMSCs cultured with different concentrations of SrCl2 revealed that the best one to improve BMSCs osteogenic differentiation was 5 μg/mL (Fig. 3). Indeed, a statistically significant increase in calcified matrix deposition was observed in cells in osteogenic medium with 5 μg/mL SrCl2, compared to osteogenic medium without SrCl2 [P < 0.01, Fig. 3(a)]. Representative micrographs showing calcified matrix stained by AR-S are reported in Figure 3(b). The osteo-inductive effect of SrCl2 was evident also on type I collagen gene expression, with cells cultured in osteogenic medium with 5 μg/mL SrCl2 showing the highest expression of this marker [P < 0.05, Fig. 3(c)]. Even ALP activity was increased by the presence of SrCl2, but differences were not significant [Fig. 3(d)]. Our results demonstrated that a low concentration of SrCl2 (5 μg/mL) improved BMSCs osteogenic differentiation, accordingly with data by Sila-Asna et al.[24] Strontium has a positive effect on bone metabolism[21] and on MSCs osteogenic differentiation.[26, 28] Furthermore, it has been demonstrated to improve osseointegration when delivered at the implant site by direct incorporation within titanium,[19] or inclusion in a hydroxyapatite coating[20] and in bone graft extenders.[46] As the in vivo effect of strontium on bone metabolism is associated with increased BMSCs osteogenesis,[27] we believe that the simultaneous inclusion of BMSCs and strontium in a hydrogel loaded into titanium may promote bone ingrowth by a dual action on host progenitor cells resident at the implant site and on cells encapsulated into the hydrogel.

Figure 3.

Selection of SrCl2 concentration to improve BMSCs osteogenic differentiation. (a) Quantification of calcified matrix after AR-S staining (ABU, arbitrary units; n = 6, ** P < 0.01). (b) Micrographs of BMSCs stained with AR-S. (c) COL1A1 expression normalized to glyceraldehyde 3-phosphate dehydrogenase (n = 3, * P < 0.05). (d) ALP activity (ALP Units/μg of proteins, n = 6). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

BMSCs culture and osteogenic differentiation in CMCA hydrogel

BMSCs viability in CMCA without titanium was determined during culture in the presence of different media: control, osteogenic, and osteogenic supplemented with 5 μg/mL SrCl2. BMSCs viability was maintained during culture in all the tested conditions [Fig. 4(a)]. ALP activity was quantified after 14 and 21 days as marker of osteogenic differentiation [Fig. 4(b)]. To confirm our data on the osteo-inductive effect of SrCl2, BMSCs were cultured in CMCA and differentiated either in normal osteogenic medium or in osteogenic medium with 5 μg/mL SrCl2. Cells cultured in CMCA maintained their osteogenic potential. Indeed, at 14 days a significant increase in ALP was observed in cells cultured in osteogenic medium with SrCl2 compared to cells in control and osteogenic medium (+ 274% and + 62% respectively, P < 0.05). The difference between this group and control cells remained statistically significant even at 21 days (+ 80%, P < 0.05). The same trend was observed in calcium levels but increases were not significant because of the great interdonor variability (data not shown).

Figure 4.

Viability and osteogenic differentiation of BMSCs on CMCA hydrogel. (a) BMSCs viability in different culture media (ABU, arbitrary units, n = 6, * P < 0.05). (b) ALP activity (ALP Units/μg of proteins, n = 4, * P < 0.05).

Accordingly with data by Leone et al.,[36] who used CMCA for chondrocyte differentiation, we did not observe any cytotoxic effect of CMCA. Culture on CMCA did not impair the ability of BMSCs to respond to the osteogenic stimuli provided by culture medium. These results, together with the already demonstrated lack of inflammatory response in rabbits treated with CMCA,[36] supports the potential clinical use of this hydrogel in bone applications.

Improvement of TT–BMSCs interaction by enrichment with CMCA hydrogel

Cell seeding efficiency and cell adhesion

The presence of CMCA inside TT pores significantly increased cell seeding efficiency [+ 21% compared to TT without CMCA, P < 0.05, Fig. 5(a)]. This increase was probably due to the ability of CMCA to immobilize cells inside TT. In view of a clinical application, it should be highlighted that the difference in cell retention between TT and TT + CMCA would be even more crucial because cells would not be allowed to adhere to TT for 3 h before getting in contact with body fluids, as we let them in our experiments before adding culture medium. We believe that prostheses made of TT preloaded with dry CMCA could be easy-to-handle devices for the orthopedic clinical practice. CMCA loaded in these devices could be hydrated intra-operatively with a suspension of autologous BMSCs obtained by bone marrow and directly implanted. After 21 days of culture, SEM showed that BMSCs colonized both TT and TT + CMCA samples, covering the surface of titanium samples [Fig. 5(b,c)] and the inner surface of pores [Fig. 5(d,e)] and displaying cell attachment to TT and TT + CMCA through the formation of cell protrusions [Fig. 5(f,g)]. These observations are coherent with recent results proving the ability of TT to support the growth of human adipose-derived stem cells.[47, 48] No difference was observed between TT and TT + CMCA demonstrating that CMCA did not impair the ability of cells to colonize and fill with matrix TT pores.

Figure 5.

Cell seeding efficiency and cell adhesion on TT and TT + CMCA. (a) Seeding efficiency expressed as percentage with respect to the initial number of seeded cells (n = 5, * P < 0.05). (b–g) SEM analysis after 21 days of culture. BMSCs adhesion on TT and TT + CMCA outer surface (b,c, scale bar 500 μm). Magnification of a titanium pore covered by cells and extracellular matrix (d,e, scale bar 200 μm). Adherent cells inside the inner surface of titanium pores (f,g, scale bar 50 μm).

BMSCs viability and osteogenic differentiation

Viability and osteogenic differentiation of BMSCs on TT and TT + CMCA were assessed to compare their osteoconductivity. As shown in Figure 6(a), no difference in cell viability was observed between TT and TT + CMCA. After 14 days, ALP activity was significantly increased in TT + CMCA samples [+ 91%, P < 0.05, Fig. 6(b)] supporting the idea that TT combined with CMCA is more osteoconductive than TT. The same trend was observed in calcium quantification, with higher values for BMSCs cultured on TT + CMCA with respect to TT [Fig. 6(c)]. Matrix characterized by a fibrillar structure was found to be deposited by cells inside titanium pores in both TT and TT + CMCA [Fig. 6(d)]. The observed increases in osteogenic markers demonstrated that the association of TT and CMCA led to a construct with improved osteoconductivity compared to TT. Further improvement of this feature by the association of TT with a hydrogel enriched with an osteo-inductive factor appears a very promising strategy to improve implant osseointegration that grants prosthetic stability.[1-3] Recently, factors such as VEGF and FGF-2 have been incorporated in hydrogel coatings to obtain bioactive titanium.[15, 16] Preliminary experiments demonstrated that SrCl2 can be included in the CMCA hydrogel and subsequently released (data not shown). In this study we found that SrCl2 at 5 μg/mL is able to improve osteogenic differentiation of BMSCs. These data will guide us in the selection of the initial amount of SrCl2 that should be incorporated into CMCA to obtain a similar final concentration released to the cells in the implant, at least in the initial phase of healing process. It is possible to speculate that, in vivo, strontium will be gradually eluted from CMCA included in a TT mesh creating a gradient of strontium concentrations that may vary with time and distance from the implant. Nevertheless, the presence of an optimal concentration of strontium within implant will provide BMSCs seeded into TT with signals triggering their initial osteogenic differentiation. In view of a future clinical application, CMCA would be enriched with SrCl2, loaded within TT mesh and oven-dried to generate off-the-shelf devices that can be easily seeded with autologous cells during surgery. In these bioactive implants, the hydrogel would play a key role in retaining progenitor cells at the bone-implant interface and delivering an osteogenic signal in addition to physiological signals from bone.

Figure 6.

BMSCs behavior on TT and TT + CMCA. (a) Cell viability of BMSCs (ABU, arbitrary units, n = 8). (b,c) ALP activity (b) and calcium quantification (c) (n = 4, * P < 0.05, ** P < 0.01). (d) Fibrillar structure of extracellular matrix produced by BMSCs after 21 days of culture (SEM analysis, scale bar 1 μm).


Faster and improved bone ingrowth into a porous titanium implant may confer better stability during the healing process. Our results demonstrate that the association of a biocompatible hydrogel, such as CMCA, with a macroporous titanium may represent a suitable approach to effectively immobilize autologous BMSCs within the implant. The enrichment with strontium and osteoprogenitor cells would convert the prosthesis from inert to bioactive, promoting osseointegration and consequently reducing the risk of mobilization and the number of revision procedures. This strategy could be exploited also in revision surgery, where macroporous titanium can be used to fill void bone spaces and to provide an anchorage point for revision components, helping the surgeon to manage bone loss at implant interface. Based on these considerations, we envision that the generation of bioactive implants that can deliver to the implant site both progenitor cells and osteogenic signals would represent a significant step forward with respect to the current clinical practice.


Authors would like to thank Stefania Pedroli, Marta Tavola, Nasser Sadr (Cell and Tissue Engineering Laboratory) and Simonetta Fusi (Limacorporate s.p.a.) for their precious help.