Received date July 11, 2012; Accepted date August 09, 2012; Published date August 11, 2012
Citation: Waselau M, Patrikoski M, Mannerström B, Raki M, Bergström K, et al.(2012) In vivo Effects of Bioactive Glass S53P4 or Beta Tricalcium Phosphate on Osteogenic Differentiation of Human Adipose Stem Cells after Incubation with BMP-2. J Stem Cell Res Ther 2:125. doi:10.4172/2157-7633.1000125
Copyright: © 2012 Waselau M, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Visit for more related articles at Journal of Stem Cell Research & Therapy
Aims: Human adipose stem cells (hASCs) have been suggested as viable alternative for bone tissue engineering. However, the tissue response and osteogenic potential of BAG S53P4 or β-TCP granules has not been studied in vivo when seeded with hASCs and/or co-incubated with BMP-2 and thus, was evaluated in the current study.
Methods and results: Human ASCs were isolated, expanded and seeded on BAG and β-TCP in vitro and, cell viability was assessed using Live/Dead staining. In a subcutaneous rodent implantation model, the cellular response and osteogenic potential of 1) plain, 2) hASC seeded, 3) BMP-2 co-incubated and 4) hASC seeded and BMP-2 co-incubated BAG and β-TCP granules were investigated using computed tomography and semi-quantitative histologic scores after 4 and 8 weeks. Live/Dead staining confirmed good cell viability on both biomaterials prior to implantation. Overall, implantation of both biomaterials resulted in formation of well-vascularized granulation tissue without excessive inflammation, fibrosis or adverse reactions independent on group assignment and time point evaluated and thus, suggesting safety for prospective applications. However, our results also indicate that β-TCP may temporarily stimulate foreign body giant cell formation after hASCs supplementation suggesting a resorptive response. Both biomaterials required supplementation of hASCs and/or BMP-2 to induce osteoblastic
activity. However, BAG induced calcification exclusively when seeded with BMP-2 activated hASCs, whereas β-TCP required seeding with hASCs only.
Conclusion: BAG and β-TCP granules can be safely implanted subcutaneously, induce a different cellular response and require hASC and/or BMP-2 supplementation to induce osteoblastic activity and calcification. A combination of β-TCP and hASCs appeared to be a feasible way in enhancing osteoblastic activity resulting in early osteogenesis while minimizing safety and regulatory concerns in bone-tissue engineering.
Bioactive glass; Beta tricalcium phosphate; BMP- 2; Human adipose stem cells; Osteogenic differentiation; In vivo; Immunocompromised rats
Healing of large bone defects after reconstructive surgery remains challenging, although technical and surgical methods have advanced rapidly over the last decades [1,2]. To enhance bone regeneration, different biomaterials and grafts have been used with variable results [1,2]. Currently, autologous bone grafts remain the gold standard for reconstructive bone surgery due to their osteoinductive and -conductive properties, low immunogenicity while offering stem cells and growth factors. However, increased donor site morbidity and limited availability  have resulted in exploration of new cell–based therapies such as bone tissue engineering .
Human adipose stem cells(hASCs) have evolved as an alternative to human bone marrow stromal cells (hBMSCs) because of simplicity in harvesting and processing while differentiating into multiple lineages such as osteoblasts . Isolated hASCs can be co-incubated with osteogenic factors to promote osteogenic differentiation in vitro  and finally, to stimulate new bone formation in vivo .
Among the biological factors used, the osteogenic potential of bone morphogenic protein 2 (BMP-2) has been well demonstrated  which resulted in approval for clinical applications [8,9]. Recently though, controversial results questioned the beneficial role of BMP- 2 on osteogenic potential of hASCs in vivo [10,11] requiring further clarification.
For bone regeneration, several biodegradable scaffold materials have been explored including bioactive glass (BAG), different calcium phosphates or polymers such as poly (lactic-co-glycolic acid) and poly- L-lactide. These polymers are mainly used as composites with more bioactive materials [12,13] since bioactivity and osteoconductivity are different as compared to BAG and calcium phosphates. However, BAG and different calcium phosphates are clinically used due to supportive effects on osteoblastic activity and proven biocompatibility [14-17]. Among the various BAG formulations used, BAG S53P4 granules have been successfully applied in reconstructive bone surgery . Similarly for calcium phosphates, beta tricalcium phosphate (β-TCP) granules gained wide access intoorthopedic surgeryincluding spinal fusions , knee surgery  and treatment of bone cysts . Although both biomaterials are routinely used, combinations of β-TCP with stem cells and/or BMP-2 have been more frequently investigated as compared to BAG [22-25]. However, our group has investigated the interaction of BAG and hASCs in the past . Both implanted biomaterials serve as initial scaffolds and cell attachment basis while inducing cell differentiation . Recently, our group has also successfully treated human patients with cranio-maxillofacial bone defects using autologous hASCs in combination with β-TCP [6,28]. Therefore, bone tissue engineered implants may represent a suitable alternative to bone grafts while containing osteogenic cells (e.g. hASCs), osteoinductive factors (e.g. BMP-2) along with a synthetic osteoconductive matrix (e.g. BAG or β-TCP). Potentially, these customized implants may unite all three bone-forming properties in a more controlled and effective manner . Although numerous studies have demonstrated the osteogenic potential of plain or growth factor and/or mesenchymal stem cell supplemented BAG and β-TCP in vitro, further in vivo studies are required prior to clinical application to confirm safety and efficacy of such customized implants. Subcutaneous implantationanimal models of bone induction are well established as initial step towards achieving these goals . Frequently, immunocompromised rats have been used for preliminary testing of xenografts to restrain immunological and excessive adverse reactions in animals [30,31] while meeting current welfare standards.
To our knowledge, the tissue response and osteogenic potential of BAG S53P4 or β-TCP granules has not been studied in vivo yet when seeded with human serum conditioned hASCs and/or co-incubated with BMP-2. In the past, these biomaterials were mainly investigated as composites in vitro, with fetal bovine serum supplemented osteogenic media or using stem cells of other species [22,26,32]. Therefore, the aim of the current study was to evaluate and compare the cellular response and osteogenic potential of BAG and β-TCP when combined with human serum conditioned hASCs and/or BMP-2 in vivo after subcutaneous implantation in immunocompromised rats. Different treatment groups were assigned as 1) plain BAG/β-TCP granules [Group 1=G 1], 2) hASC seeded [G 2], 3) BMP-2 co-incubated [G 3] and 4) hASC seeded and BMP-2 co-incubated [G 4] BAG/β-TCP granules. Our hypothesis was that BAG and β-TCP granules can 1) be safely implanted subcutaneously, 2) induce a different cellular response and 3) require hASC and/or BMP-2 supplementation to induce osteoblastic activity and calcification.
Ethics statement, hASCs isolation and culture
Harvesting and isolation of hASCs was conducted in accordance with the Declaration of Helsinki and the Ethics Committee of the Pirkanmaa Hospital District, Tampere, Finland (R03058). Human ASCs were isolated from adipose tissue samples collected from three female donors (age 41 ± 8 years) undergoing elective surgical procedures in the Department of Plastic Surgery, Tampere University Hospital (Tampere, Finland). Isolation of ASCs was carried out using a mechanical and enzymatic method as described previously [26,33]. Thereafter, isolated cells were maintained and expanded in polystyrene flasks (Nunc, Roskilde, Denmark) in medium containing Dulbeccos’s modified Eagle medium/Ham’s nutrient mixture F-12 (DMEM/F-12 1:1; Invitrogen, Germany) which was supplemented with 1% L-glutamine (GlutaMAX; Invitrogen), 1% antibiotics (100U/ ml penicillin, 0.1mg/ml streptomycin; Invitrogen) and 10% human serum (PAA Laboratories GmbH, Pasching, Austria) at 37°C and 5% CO2. When cells reached about 80% confluency, they were passaged and detached enzymatically with trypsin (TrypLE SelectTM, Invitrogen). Before implantation, hASCs were expanded, and cells isolated from three patients were pooled. Cell passages 2-3 were used for cell seeding on biomaterials. The genomic stability of hASCs was verified using standard chromosome G-banding method. Karyotypic evaluation was performed as an outsourcing service in Medix Laboratoriot Ltd, Espoo, Finland.
Flow cytometric analysis of hASC surface marker expression
After primary culture for 1-2 cell passages, hASCs (n=3) were harvested and characterized by flow cytometry (FACSAriaTM; BD Biosciences, Erembodegem, Belgium) as described previously . Cell samples with 100,000 cells were single-stained with monoclonal antibodies against CD11a-allophycocyanin (APC) (1:10), CD80- phycoerythrin (PE) (1:10), CD86-PE (1:10), CD105-PE (1:10) (R&D Systems Inc., Minneapolis, MN, USA), CD3-PE (1:10), CD14- phycoerythrin-Cyanine (PECy7) (1:50), CD19-PECy7 (1:25), CD45RO-APC (1:30), CD54-fluorescein isothiocyanate (FITC) (1:10), CD73-PE (1:10), CD90-APC (1:200) (BD Biosciences), and CD34- APC (1:10), HLADR-PE (1:10) (Immunotools GmbH, Friesoythe, Germany) were used. Analysis was performed on 10,000 cells per sample and unstained cell samples were used to compensate for the background autofluorescence.
Biomaterials and growth factor
Clinically used biomaterials were selected for the study. Bioactive glass granules (S53P4, 23% Na2O, 20% CaO, 53% SiO2, 4% P2O5; BonAlive granules, 1.0-2.0mm; BonAlive Biomaterials Ltd., Turku, Finland) and beta tricalcium phosphate granules (ChronOS granules, 1.4-2.8mm, porosity 60%; Synthes, Oberndorf, Switzerland) were used in the study. Also, clinically used BMP-2 (Sigma-Aldrich®, St. Louis, MO, USA) was chosen as growth factor supplementation for osteogenic medium.
Seeding and osteogenic differentiation of hASCs on biomaterials
Bioactive granules were sterilized using 70% EtOH prior to cell seeding. Briefly, BAG was washed twice with sterile 70% EtOH followed by incubation in 70% EtOH for 1.5 hour. Subsequently, 70% EtOH was removed and BAG was dried under a laminar flow for 50 minutes. Finally, BAG was rinsed six times with DBPS to completely eliminate EtOH. Sterilization of β-TCP granules prior to cell seeding was not required since the biomaterial was pre packed aseptically. Equal volumes of sterile BAG and β-TCP granules (granule volume = 400μl) were placed into each well of a 24-well plate (Nunc, Roskilde, DK- 4000) followed by pre-incubation with medium for 48 hours before cell seeding as described earlier . Subsequently, both biomaterials were treated resulting in the following combinations: 1) BAG only, 2) BAG + hASCs 3) BAG + BMP-2, 4) BAG + hASCs + BMP-2 5) β-TCP only, 6) β-TCP + hASCs, 7) β-TCP + BMP-2, 8) β-TCP + hASCs + BMP- 2 (Table 1). Biomaterials were seeded with previously isolated hASCs using 210,000 cells/well  containing 70,000 cells from each donor and osteogenic differentiation was induced with medium containing exogenously supplemented BMP-2 (200 ng/ml) as described earlier [11,22,35]. Four hundred μl of seeded grafts were maintained in culture for seven days at 37.5°C and 5% CO2 changing individual media every 48 hours until implantation as described elsewhere .
|Group||Surgery at wk 0||Harvesting at wk 4||Harvesting at wk 8|
|G11 = BAG2||8 rats||4 rats||4 rats|
|G 2 = BAG + hASCs||8 rats||4 rats||4 rats|
|G 3 = BAG + BMP-2||8 rats||4 rats||4 rats|
|G 4 = BAG + hASCs + BMP-2||8 rats||4 rats||4 rats|
|G 1 = b-TCP2||8 rats||4 rats||4 rats|
|G 2 = b-TCP + hASCs||8 rats||4 rats||4 rats|
|G 3 = b-TCP + BMP-2||8 rats||4 rats||4 rats|
|G 4 = b-TCP + hASCs + BMP-2||8 rats||4 rats||4 rats|
1G=group, 2Represents control group for each biomaterial
Table 1: Randomly assigned groups.
Qualitatively, viability of hASCs was evaluated using Live/Dead staining (Molecular probes, Eugene, OR, USA) 5 days after cell seeding as described elsewhere . Briefly, samples were evaluated with a fluorescence microscope after incubation in 5μM CellTrackerTM green (5-chloromethylfluorescein diacetate [CMFDA]; Molecular Probes) and 2.5 μM Ethidium Homodimer-1 (EH-1; Molecular Probes) for 45 minutes. With this method, viable cells stained green whereas red fluorescence indicated dead cells.
Sixty-four (64), 10 weeks old immunodeficient, athymic, male RNU rats (Crl:NIH-Foxn1rnu, strain code:316; Charles River Laboratories International Inc., Wilmington, MA, USA) were used in the current study. The animal study was performed according to national laws of animal welfare and protection with the permission of the local Ethical committee and official veterinary authority (animal research permit number ESAVI/664/04.10.03/2011). The study was in compliance Animal Protection Act (§7,8,9; EU-convention on the protection of animals revised directive 86/609/EEC). In pairs (2), all animals were allowed to acclimatize in individually ventilated cages preoperatively for 14 days. Their general health status was assessed twice daily and free access to sterilized water and peletted food was allowed until surgery.
Eight (8) animals were randomly assigned to each of the following treatment groups: group 1 = BAG or β-TCP (control group = CG); group 2 = BAG or β-TCP+ hASCs; group 3 = BAG or β-TCP + BMP-2 and group 4 = BAG or β-TCP + hASCs + BMP-2. Four rats from each treatment group and biomaterial were sacrified for analyses after 4 and 8 weeks (Table 1).
Anesthesia, surgical procedure and postoperative care
Thirty minutes prior to surgery, all animals were evaluated to ensure that their general health status was within normal limits followed by administration of pain medication (carprofen, 5mg/kg s.c., Pfizer GmbH, Berlin, Germany) and antibiotics (enrofloxacine, 10mg/kg i.m. once; Bayer AG, Leverkusen, Germany). Animals were transferred from their cages to the surgery table under a laminar flow. Anesthesia was induced using isoflurane at 4-5% and maintained with 2-5% in 100% oxygen delivered in a semi-closed system. Thereafter, animals were positioned in sternal recumbency, their vital status was continously monitored and, proposed implantation sites were shaved and aseptically prepared using chlorhexidine solution and alcohol. Subsequently, a sterile adhesive drape was used to cover the proposed surgical site. Two 2-cm long skin incisions were centered 2cm parallel to, and on either side of the spine using # 15 surgical blade. Subsequently, the subcutaneous tissue was slightly and bluntly undermined using Mosquito hemostatic forceps. Care was taken to avoid excessive tissue damage, hemorrhage and to allow sterile placement of the biomaterials/sterile conglomerates into each cavity according to the predetermined and randomly assigned schedule. Subsequently, skin edges were reapposed in a simple interrupted suture pattern using # 3-0 prolene. After recovery from anesthesia, animals were returned to their individual predetermined cages. Postoperatively, all animals received pain medication for 3 days (carprofen, 5mg/kg s.c. q24h, Pfizer GmbH). Twice daily, their general health status, water/food intake, lameness, surgical site and urination/defecation was monitored and scored until termination of the study at 4 and 8 weeks when rats were euthanized using the carbon dioxide method. Additionally, animals were weighted once weekly throughout the study period.
Computed tomographic (CT) imaging
Immediately after sacrificing, rats were placed into a CT-scanner to evaluate presence and measure volume of implanted biomaterials. CT imaging was performed with dedicated small animal scanner nanoSPECT/CT (Bioscan Inc, Washington DC, USA) featuring conebeam CT system at 100 μm resolution. Circular CT scans were acquired in 360 projections at 55 kVp tube peak voltage by using exposure time of 1500 ms. After imaging, datasets were reconstructed and analyzed with In Vivo Scope software (Bioscan Inc.). Three-dimensional volumes of interest (VOIs) were defined around biomaterial regions by using isocontour tool. Six slices were analyzed at a time and volumes were summed together in order to obtain total volume of the implant (mm3).
Sample harvesting and histology
After computed tomographic imaging, implanted biomaterial conglomerates were isolated under sterile conditions leaving granules subcutaneously attached. All specimens were evaluated macroscopically and photographed. Subsequently, samples were processed as described previously . For 5-7 days, specimens were fixed in 4% formalin solution, washed in phosphate buffered saline and dehydrated in series of ascending ethanol concentrations. Subsequently, samples were defatted in xylene, infiltrated with methyl methacrylate (methacrylacid– methylester; dibuthylphtalate and perkadox in a 89,5:10:0,5 proportion) and embedded in the same solution using Teflon–molds in a water bath at 30°C. Ground sections (30 μm) were cut as reported earlier  and stained with toluidine blue (TB).
Histological parameters were semi-quantitatively scored as described elsewhere [39,40]. The presence (score 0=none, 1=present) and thickness of periimplantar capsules were evaluated (capsule thickness=CT; score: 0=none present; 1=minimal/ < 4 cell layers; 2=mild/ < 8 cell layers; 3=moderate/< 12 cell layers; 4=marked/> 12 cell layers). Also, the degree of granulation tissue  formation between BAG and β-TCP granules was evaluated (scores: 0=none visible; 1=minimal/< 25% of granules enclosed by GT per visible power field; 2=mild/< 50% of granules enclosed; 3=moderate/< 75% of granules enclosed; 4=marked/> 75% of granules enclosed). The degree of vascularization of the GT was scored accordingly (scores: 0=no vessels visible per power field; 1=minimal/< 4 vessels; 2=mild/< 8 vessels, 3=moderate/< 12 vessels; 4=marked/> 12 vessels). Further, immunological and inflammatory reactions within the periimplantar GT were evaluated for cell density (CD), macrophages (M?), foreign body giant cells (FBGC), spindle–shaped cells (SSCs), amount of osteoblastic precursor cells (OBPCs) (scores: 0=no cells per visible power field; 1=minimal/< 20 cells; 2=mild/< 40 cells; 3=moderate/< 60 cells; 4=marked/> 60 cells) and degree of calcification (score: 0=no calcification per visible power field; 1=minimal/< 20% calcification; 2=mild/<40% calcification; 3=moderate/< 60% calcification; 4=marked/> 60% calcification). For the β-TCP group exclusively, fillings of granule pores were evaluated (score: 0=no filling per visible power field; 1=minimal/< 20% filling; 2=mild/<40% filling; 3=moderate/< 60% filling; 4=marked/> 60% filling). The vascularization of pore filling was scored accordingly (scores: 0=no vessels per visible power field; 1=minimal/< 4 vessels; 2=mild/< 8 vessels; 3=moderate/< 12 vessels; 4=marked/> 12 vessels). All above-mentioned parameters were scored semi-quantitatively by two blinded reviewers. If results differed, mean values were calculated for statistical analyses. Results of each group were compared among and between each other.
A Kruskal-Wallis test with Dunn’s multiple comparison test was used for the statistical analysis using Prism® (GraphPad Prism 5.01 Software, Inc., La Jolla, CA, USA) software and were presented as mean and standard deviation (SD). The level of significance was set at level p < 0.05.
Isolation and characterization of hASCs
Human ASCs were successfully isolated and expanded using cell culture medium containing 10% human serum. Flow cytometric analyses confirmed expression of surface markers characteristic for hASCs. Analyses showed positive expression (>80%) for the markers CD73, CD90, CD105 and lacked (<2%) the expression of the markers CD11a, CD14, CD19, CD45 and HLA-DR (Figure 1). Moderate expression was observed for the hematopoietic progenitor marker CD34, however, the variance between samples was high. In addition, cells lacked (<2%) the expression of immunogenicityassociated markers CD3, CD80 and CD86 and demonstrated moderate expression of the immune and endothelial cell marker CD54 (Figure 1). The expression pattern was typical for hASCs .
Figure 1: Overview on FACS-Results
The surface marker expression was characteristic for undifferentiated hASC as analyzed by flow cytometry. The graph shows the fluorescence intensity (x-axis) and the relative cell number (y-axis). Unstained control cells are represented as empty histograms, whereas antibody stained cells are presented as filled histograms. The unstained control sample dot plot is showing particle size and granularity (side scatter vs. forward scatter).
Live-Dead stainings revealed viability of hASCs on surfaces of BAG and β-TCP (Figure 2) prior to implantation suggesting a supportive effect of human serum containing medium on viability. BMP-2 supplementation did not have an adverse effect on hASCs viability.
Figure 2: Live/Dead stainings at day 5
Human adipose stem cells were successfully seeded on BAG and β-TCP granules and the majority remained viable prior to implantation. Green fluorescence confirmed viability of cells, whereas dead cells stained red. Unfocused spots indicate different layers of biomaterials due to 3-D structure and reflective surfaces of the material.
Anesthesia, surgical procedure and postoperative care
All rats were successfully operated without intra- or postoperative complications. The subcutaneous implantation model was well tolerated by all animals and no adverse reactions or negative effects on their wellbeing were observed at any time during the study. Animals survived as planned until termination of the study at week 4 and 8.
Computed tomography imaging
Computed tomography evaluation confirmed proper subcutaneous implantation of BAG (Figure 3) in all groups. Generally, granules of both biomaterials were densely packed and remained at the original surgical site at week 4 and 8. Volume analyses revealed no differences between surgical sites for either biomaterial in any group (Vright=237.63mm3, SD=27.2; Vleft=236.56mm3, SD=26.1; p=0.822).
Figure 3: Postoperative computed tomographic images
The figure shows representative images for implanted BAG and β-TCP granules at week 8. Dorsoventral and transversal views on computed tomographic imaging confirmed standardized surgical procedures at week 4 and 8, since no volume differences were measured between surgical sites. Furthermore, biomaterials remained at the original implantation sites over time.
Similarly for both biomaterials, the volumes decreased or remained the same over time except for group 4 in BAG groups, where the volume was significantly larger at week 8 than at week 4 (p=0.036) (Table 2). Further, similar volumes were measured over time among β-TCP and BAG groups, except for group 3 as compared to group 4 (p=0.03) among BAG groups at week 4. When compared to β-TCP, the volume of BAG was not different in any group at week 4. However, at week 8, the volumes of BAG group 4 samples were larger than β-TCP group 4 samples (p=0.005).
|G1 1 = BAG2||4||248.540||14.747||0.522|
|G 2 = BAG + hASCs||4||257.867||16.843||0.028|
|G 3 = BAG + BMP-2||4||268.229||18.586||0.030|
|G 4 = BAG + hASCs + BMP-2||4||234.188||14.845||0.036|
|G 1 = TCP2||4||228.338||19.438||0.764|
|G 2 = TCP + hASCs||4||215.266||34.898||0.789|
|G 3 = TCP + BMP-2||4||256.493||25.076||0.017|
|G 4 = TCP + hASCs + BMP-2||4||229.011||11.586||0.015|
1G=group, 2Represents control group for each biomaterial
Table 2: Week Comparison of Biomaterial Volumes at Surgical Sites on CT.
Sample harvesting and macroscopical evaluation
All implanted biomaterial conglomerates were successfully isolated. Subsequent inspection revealed no foreign body or excessive inflammatory reactions, but a thin and translucent capsule surrounded all specimens. Furthermore, small vessels were arising from the subcutaneous tissue and were merging centripetally through capsules. These vessels provided vascularization of the periimplantar granulation tissue as observed microscopically (Figure 4).
Figure 4: Representative images of external capsule formation and vascularization of periimplantar granulation tissue. The figure shows representative images for implanted BAG and β-TCP granules at week 8. At weeks 4 and 8, samples were harvested and demonstrated similar macroscopical appearance independent on biomaterial implanted. Vessels (BAG/β-TCP – A; black arrows) were sprouting centripetally into all capsules indicating viability. A translucent and thin capsule (BAG/β-TCP – B; red arrows) formed around all specimens. Infiltrating vessels provided vascularization of periimplantar granulation tissue (BAG/β-TCP – C; green arrows). Asterisks indicate biomaterials.
An overview on all semi-quantitative histological scorings is given in Figures 5, 6 and 7. Overall, histologic evaluation revealed no differences between the BAG and β-TCP group regarding capsule thickness, amount of periimplantar granulation tissue degree of vascularization, cell density, spindle-shaped cell or macrophage accumulation (Figure 5 and 6). However in group 2, significantly higher accumulation of foreign body giant cells was observed at week 4 in β-TCP group as compared to BAG group (Figure 6). In β-TCP group, significantly higher scores for osteoblastic precursor cell formation were recorded in groups 3 and 4 as compared to group 1 (Figure 7). Additionally, more periimplantar calcifications were noticed around β-TCP granules in group 2 at week 8 as compared to BAG (Figure 7 and 8).
Figure 5: Overview on semi-quantitative scoring of histology sections The figure shows representative images for BAG and β-TCP at week 8. At weeks 4 and 8, no statistical differences were observed among treatment groups and time points evaluated for capsule thickness (CT), cell densities (CD), spindle-shaped cells (SSCs), amount of periimplantar granulation tissue (GT) and vascularization. G1 – G4 = groups 1-4.
Figure 6: Overview on semi-quantitative scoring of macrophage and foreign body giant cell activity around biomaterial granules Overall, similar macrophage activities were observed around BAG and β-TCP granules. However, temporarily increased foreign body giant cell (FBGCs) accumulation was observed after implantation of hASC seeded β-TCP granules as compared to BAG granules at week 4. G1 – G4 = groups 1 - 4; arrows with asterisks indicates statistical differences between BAG and β-TCP.
Figure 7: Overview on semi-quantitative scoring of osteoblastic precursor cell formation and calcification around BAG and β-TCP granules Subcutaneous implantation of plain BAG or β-TCP granules failed to induce osteoblastic activity and calcification, but seeding of both biomaterials with hASCs and/or supplementation with BMP-2 induced osteoblastic activity. However, calcification was only observed, when BMP-2 activated hASCs were seeded on BAG or hASCs were seeded on β-TCP. G1 – G4 = groups 1 – 4; OBPCs = osteoblastic precursor cells; d = statistical difference between groups 2 and 3; e = statistical difference between groups 2 and 4; f = statistical difference between groups 1 and 4; N.D. = not detected (score = 0); single asterisks indicates statistical differences between weeks 4 and 8; arrows with asterisks indicates statistical differences between BAGand β-TCP.
Figure 8: Histological sections of periimplantar tissue after implantation of BAG and β-TCP granules
The figure shows representative images for BAG and β-TCP at week 8. BAG – A = healthy, vascularized granulation tissue was formed around all granules containing spindle-shaped cells (red arrow); BAG – B = macrophages (green arrow) and foreign body giant cells (red arrow) were simultaneously observed; β-TCP – A = typical appearance of foreign body giant cells (red arrow) accumulating at the surface of a β-TCP granule; β-TCP – B = preosteoblasticcell nests (green arrow) were only observed after seeding with hASCs and/ or BMP-2 supplementation, but early calcification along β-TCP granules (red arrow), was only observed after hASC seeding. Asterisks represent biomaterials.
Bone tissue engineering may prospectively substitute bone grafts, but more in vivo studies are required to prove safety and efficacy prior to clinical application . To our knowledge, this is the first in vivo study evaluating and comparing both, the tissue response and osteogenic potential of BAG and β-TCP when seeded with human serum conditioned hASCs and/or co-incubated with BMP-2 in a subcutaneous model using immunocompromised rats.
Human ASCs represent an alternative stem cell source for bone tissue engineering because of abundant availability, greater stem cell yield, low immunogenicity and similar osteogenic differentiation potential as compared to hBMSCs [44,45], which may simplify clinical application and thus, were evaluated in the present study. Patient safety is of major concern in clinical application of bone tissue engineered products, requiring the elimination of risk factors. Traditionally, mesenchymal stem cells have been expanded and differentiated in medium supplemented with fetal bovine serum. However, animal derived components are a concern for clinical application in humans. Therefore, human serum has been investigated as a potential substitute [34,46] to overcome those safety concerns [47,48] and it has been successfully used along with hASCs and biomaterials to reconstruct large maxillary and cranial bone defects in patients . Our results confirmed recent reports in which hASCs were successfully isolated and expanded in cell culture medium supplemented with human serum [6,49,50]. Additionally, our flow cytometric analyses confirmed characteristic immunophenotype of hASCs, which was similar to earlier reports [5,33,42]. Live/Dead staining confirmed good cell viability on both biomaterials prior to implantation. Our results were in alignment with earlier reports in which viability of hASCs on BAG and β-TCP surfaces had been demonstrated [28,36].
Computed tomography is as a sensitive tool to evaluate presence and volume of implanted biomaterials  and was thus used in the current study. Overall, our data confirmed a robust, standardized surgical procedure since implanted biomaterial remained at the surgical site over time and postoperative volume was similar.
In general, tissue response of implants can be histologically evaluated by assessing foreign body and inflammatory reactions within the periimplantar tissue . In the current study, these parameters were evaluated at week 4 and 8 allowing simultaneous assessment of implant related immediate-term adverse reactions and the osteogenic potential while minimizing the animal number for welfare reasons. The biocompatibility of BAG  and β-TCP  was demonstrated in previous studies. Consistent with our study, external capsule formation secondary to subcutaneous implantation of both biomaterials was recently described [54,55]. However, this tissue reaction can be considered as a physiological and mechanical response , since similar cell densities were recorded for both biomaterials and excessive fibrosis (SSCs), inflammatory cell accumulation (macrophages) or avascularity was not observed in the current study. The minimal fibrotic potential of both biomaterials may be well related to mechanical friction since BAG and β-TCP may crumble slightly and may, therefore, increase resistance with subsequent capsule formation as reported earlier . However, supplementation of BAG and β-TCP with BMP-2 and/or hASCs did not result in excessive capsule formation or fibrosis suggesting equally favorable tissue response and safety of the combinations for prospective applications. This was also supported with the equally low number of inflammatory cells and with the viable granulation tissue, which was infiltrated by centripetally merging vessels in the current study.
Vascularized granulation tissue is formed as temporary space holder in bone healing  and the supportive effects of BAG  and β-TCP  on granulation tissue formation have been demonstrated. Similar to these reports, both biomaterials induced a moderate amount of vascularized granulation tissue over time supporting prospective healing. Interestingly, concurrent application of BMP-2 and/or hASC had no enhancing effect on granulation tissue formation independent on time and biomaterial. This is in contrast to previous studies, demonstrating faster granulation and secondary bone healing when rhBMP-2 was concurrently applied with either BAG  or β-TCP [62,63] into osseous defects. However, the discrepancy may be related to the BMP-2 delivery method and the heterogenous implantation model in the current study but requires further clarification.
Vascularization is essential for bone healing  and the positive effects of BAG  and β-TCP  on angiogenesis have been proven. In alignment with those studies, our data confirmed formation of vascularization after implantation of both biomaterials but without superior effect of one biomaterial over the other. Addition of hASCs and/or BMP-2 did not further enhance vascularization. Furthermore, vascularization of β-TCP granule pores was observed independent on treatment group in the present study, indicating good tissue viability within the β-TCP granules. The lack of a proangiogenic potential of hASCs and/or BMP-2 supplementation is in contrast to a previous report demonstrating increased vascularization in apatite-coated scaffolds when seeded with hASCs and stimulated with BMP-2  and requires further investigations. Overall our data indicate no negative effects on vascularity of hASCs and/or BMP-2 supplementation of BAG or β-TCP granules in the current study demonstrating safety of prospective application.
The tissue response to BAG  and β-TCP  has been documented in vivo based on macrophage and foreign body giant cells counts. Overall, similar activation of macrophages after application of both biomaterials independent on hASCs and/or BMP-2 application was observed in the current study suggesting good biocompatibility without excessive inflammation. Our data were in contrast to previous studies in which greater macrophage accumulation was observed after BAG application as compared to β-TCP when either implanted plain or when supplemented with BMP-2 [70,71]. Further, the macrophage counts neither increased over time nor were significantly increased after supplementation with BMP-2 or seeding with BMP-2 co-incubated hASCs as observed in a recent in vivo study demonstrating a dosedependent inflammatory response of rhBMP-2 after subcutaneous and intramuscular injections . Interestingly, β-TCP appeared to be pro-inflammatory when seeded with hASCs at week 4 as compared to BAG since increased foreign body giant cell activity was recorded. This is in contrast to recent studies [73,74], suggesting antiinflammatory and immunomodulatory effects of mesenchymal stem cells. However, similarly to our results, β-TCP attracted significantly greater multinucleated giant cell formations within the subcutaneous implantation bed and thus, may be considered as physiological degradation response in vivo . Further, the tissue response has to be also graded according to the nature of the material implanted. In contrast to BAG, β-TCP is more rapidly degraded and thus, an increased number of multinucleated giant cells can be anticipated . Overall though, foreign body giant cell formation was minimal when both biomaterials were co-incubated with hASC and/or BMP-2 suggesting safety of the combination.
Osteoblastic cells and calcification are indicators for in vivo osteogenesis . In the current study, application of plain BAG or β-TCP failed to induce either osteoblastic activity or calcification within 8 weeks. Furthermore, the lack of substantial periimplantar bone formation on histological samples was also proved with computed tomography. Our observation may be related to the subcutaneous implantation model since bone formation was observed as early as 4 weeks in other studies after intramuscularor interosseous application of BAG  or β-TCP compositions . Similarly, simple BMP-2 supplementation of both biomaterials failed to induce calcification in the present study, although the osteogenic potential of BMP-2 has been documented [79,80] and may be related to different dosages and application methods. However, seeding of both biomaterials with hASCs and/or co-incubation with BMP-2 induced osteoblastic precursor cell formation over time confirming their pivotal role in osteogenesis [36,81]. In the current study, supplementation of BAG with hASCs and/or BMP-2 was equally effective in osteoblastic precursor cell induction suggesting similar osteoblastic potential of all combinations but none of the latter combinations were superior when compared to β-TCP indicating similar osteogenic potential. However, only a combination of BAG with BMP-2 and hASCs resulted in calcification over time indicating a synergistic effect of both factors on osteogenic differentiation and osteogenesis. Potentially, BMP-2 induced osteogenic differentiation of hASCs and host derived cell migration as well as paracrine angiogenic factors from seeded hASCs contributed to osteogenesis. However, BAG granules are also more rigid as compared to β-TCP  resulting in mechanical stimulation. The latter has been suggested to promote and maintain calcification [83,84]. Our data are in contrast to previous studies, in which BMP-2  and mesenchymal stem cells  had positive effects on osteogenesis when individually incubated with BAG for 8 weeks although these studies were performed using BMP-2 gene transfer and modified scaffolds. Therefore, our conflicting results regarding calcification potential may be related to specific biomaterial composition, type of stem cell, BMP- 2 delivery method and/or implantation model used. Similarly, β-TCP initiated osteoblastic precursor cell activity when seeded with hASCs and/or BMP-2 in the current study. However, BMP-2 supplementation or BMP-2 activated hASCs induce significantly higher osteoblastic activity as compared to hASCs alone suggesting a greater potential of the latter combinations. This is in contrast to recent in vitro studies [11,35,85] suggesting that osteogenic differentiation of hASCs is not affected by BMP-2 and thus, may not be a viable strategy for bony healing. Furthermore, seeding of β-TCP alone with hASCs resulted in calcification in the current study, which was significantly greater as compared to BAG group confirming a clinical case series, in which similar composites sealed large craniotomy defects . Potentially, paracrine factor secretion from seeded hASCs may have contributed to calcificaton in the current study. However recently, the value of these composites was questioned since even higher densities of seeded hASCs failed to form bone in β-TCP scaffolds emphasizing the need for osteogenic factors to support complete osteoblastic cell differentiation in vitro with subsequent osteogenesis in vivo  and thus, further clarification is needed. Also in contrast to our results, ectopic bone formation of BMP-2 supplemented calcium phosphate scaffolds was demonstrated in rats after 8 weeks . Similarly, β-TCP/fibrin glue composites seeded with hBMSCs resulted in ectopic osteogenesis after 8 weeks . However, the different response in the current study may reflect differences in biomaterial composition, type of stem cells and implantation model used.
In summary, although no adverse reactions were observed and vascularized granulation tissue was formed, different cellular responses were induced in vivo. Beta tricalcium phosphate stimulated foreign body giant cell formation after hASCs supplementation early on indicating a resorptive response. Both biomaterials required supplementation of hASCs and/or BMP-2 to induce osteoblastic activity. Bioactive glass induced calcification exclusively when seeded with BMP-2 activated hASCs, whereas β-TCP required seeding with hASCs only.
Human ASCs can be successfully isolated and expanded in medium containing human serum followed by seeding on BAG and β-TCP granules. Our data suggest that both biomaterials can be safely used in combination with hASCs and BMP-2 when subcutaneously implanted. Based on our results, a combination of β-TCP and hASCs appeared to be a feasible way in enhancing osteoblastic activity resulting in osteogenesis while minimizing safety and regulatory concerns in bone tissue engineering. Although the current study suggested safety for application, further in vivo studies using interosseous implantation models are required to prove biocompatibility and efficacy. However, this study was a preliminary step towards achieving this goal.
The work was supported by TEKES, the Finnish Funding Agency for Technology and Innovation as well as the competitive research funding of the Pirkanmaa Hospital District in Finland (9M058, 9N042). The authors would like to thank Käthi Kämpf, Sabina Wunderlin and Katalin Zlinszky for processing of histology samples and Anna-Maija Honkala, Miia Juntunen and Sari Kalliokoski for assisting with the in vitro part of the study.