The aim of this pilot study was to determine, in a new experimental model, whether complex bioartificial monoblocs of relevant size and stability can be prefabricated in a defined three-dimensional design, in which the latissimus dorsi muscle serves as a natural bioreactor and the thoracodorsal vessel tree is prepared for axial construct perfusion. Eighteen sheep were included in the study, with six animals in each of three experimental groups. Vitalization of the β-tricalcium phosphate-based constructs was performed by direct application of unmodified osteogenic material from the iliac crest (group A), in vivo application of nucleated cell concentrate (NCC) from bone marrow aspirate (group B), and in vitro cultivation of bone marrow stromal cells (BMSC) in a perfusion bioreactor system (group C). The contours of the constructs were designed digitally and transferred onto the bioartificial bone grafts using a titanium cage, which was bent over a stereolithographic model of the defined subvolume intraoperatively. At the end of the prefabrication process, only the axial vascularized constructs of group A demonstrated vital bone formation with considerable stability. In groups B and C, the applied techniques were not able to induce ectopic bone formation. The presented computer-assisted workflow allows the prefabrication of custom-made bioartificial transplants.
The transplantation of autogenous bone grafts is still the gold standard for reconstruction of bony defects in the facial skeleton after trauma, infection, or tumour ablation. Depending on the size of the transplant and the soft tissue situation at the site of transplantation, autogenous bone grafts can be harvested as vascularized bone grafts for microvascular tissue transfer, or as non-vascularized bone grafts for free transplantation. In the facial skeleton in particular, there is the need for an adequate biological reconstruction, since bone grafts can be exposed directly to the paranasal sinuses or can communicate indirectly with their environment via dental implants. In situations with compromised perfusion due to radiation or infection, microsurgical bone transfer is preferred for osseous reconstruction, inducing a healing process from the transplanted graft comparable to bone fracture healing. However, the availability of autogenous bone grafts is limited, and their preparation can be associated with considerable morbidity at the donor site, which also seems to correlate with the amount of harvested bone. Furthermore, the absolute size and three-dimensional (3D) shape of autogenous bone grafts often compromise the reconstruction of the true original facial structures, which should aim for functional and aesthetic recovery. The bony frame of the face serves equally for support of the facial contours and global position and muscle insertion for mastication and facial expression, as well as incorporation of dental implants in the upper and lower jaws, emphasizing the need for an innovative approach to create an individual and biologically adequate solution for surgical reconstruction in this complex anatomic region.
Non-vascularized allogeneic and xenogeneic transplants are also difficult to adapt to an individual defect situation and in addition are associated with the risk of disease transmission. Although osteoinductive properties can be preserved to some degree by modern processing techniques, higher complication rates after free transplantation seem to indicate a lower osteogenic potential compared to autogenous transplants.
Different types of innovative biomaterials with excellent osteoconductive properties are also available for surgical reconstruction. In an experimental study, Schliephake et al. could evoke bone ingrowth into a porous matrix in direct contact with the underlying mandibular bone under guided bone regeneration with resorbable membranes at the ascending ramus, and could later transplant these constructs into the horizontal ramus as onlay grafts. Absorbable as well as non-absorbable bone substitutes of biological and synthetic origin are frequently used for refilling of small bony defects, but show a lack of ossification and stability when they are used for large bony defects. Therefore, several approaches have been made in order to raise the number of osteogenic cells on scaffolds, including intraoperative procedures for concentration of bone marrow-derived stem cells, as well as modern tissue engineering techniques with in vitro cell cultivation in perfusion bioreactor systems, which are used to raise bone in 3D scaffolds before transplantation. Regarding these techniques, bone growth still seems to be reduced to the surface of the scaffolds due to the limited supply of oxygen and nutrition inside. There are numerous studies reported in the literature that have examined bone growth on scaffolds in vitro, or sometimes in small animals, but the number of experimental studies with in vitro seeded scaffolds of clinically relevant size in large animals, or even clinical cases in humans, are rare.
The current challenge in this field is the induction of angiogenesis for continued ossification after implantation of seeded scaffolds into the in vivo environment.
The focus of research has been on different vascularization techniques in order to improve the blood supply inside seeded scaffolds, including the surgical implementation of vessel loops or arteriovenous vascular bundles. Although the activated mechanisms and complex cascades of interactions are still not known in detail, surgical angiogenesis appears to be an elementary requirement for successful prefabrication of bioartificial constructs as needed for surgical reconstruction. Previous studies have already shown that bone growth, ceramic resorption, and angiogenesis increase significantly with axial perfusion of bioartificial constructs, especially when autologous osteogenic material from the iliac crest is used directly for construct vitalization.
The aim of the present pilot study was to determine whether bioartificial monoblocs of relevant size and stability can be generated in a defined 3D design within a prefabrication process, in which the thoracodorsal vessel tree is prepared for axial construct perfusion and could potentially later serve for microvascular anastomosis. In order to overcome the limitations of in vitro tissue engineering techniques, autologous cancellous bone and amorphous bone marrow (BMA) from the iliac crest were harvested for construct vitalization and directly implanted into the natural bioreactor environment of the latissimus dorsi muscle. The effectiveness of this convenient procedure was compared with other more expensive procedures for construct vitalization with special regard to their suitability for clinical use. The 3D shape of the complex constructs was designed digitally on the computer and transferred onto the bioartificial bone grafts using the outer contour of a titanium cage, which was bent over a stereolithographic model of the defined subvolume intraoperatively. The described workflow could principally serve for prefabrication of custom-made bioartificial transplants. Furthermore, the titanium protected thoracodorsal vessel tree in the latissimus dorsi muscle is introduced as a new experimental model, which could also serve for other preclinical studies on large animals with axial vascularized scaffolds in a natural bioreactor environment, independent of construct configuration, chosen material, or applied seeding procedure.
According to the literature, a considerable number of research groups are working on the development of new bone grafting materials, carriers, growth factors, and specifically tissue-engineered constructs for bone regeneration. To allow comparison between different studies and their outcomes, it is essential that animal models, fixation devices, surgical procedures, and methods of taking measurements are well standardized to produce reliable data pools. Animal models in bone repair research already include representations of normal fracture healing and fracture non-unions, as well as critical-sized segmental bone defects. However, an experimental model for ectopic en bloc prefabrication of vascularized bioartificial bone grafts for preclinical studies on large animals – as clearly needed – has not been established yet. The titanium protected thoracodorsal vessel tree in the sheep meets all the requirements for a recognized preclinical experimental model, including physiological and pathophysiological analogies to humans in respect to the scientific question, simple surgical execution, possible observation of a multiplicity of study objects, costs for acquisition and care, animal availability, acceptability to society, tolerance to captivity, and ease of housing.
Materials and methods
Animal experiments were conducted under a protocol approved by the ethics committee in accordance with German federal animal welfare legislation. Eighteen healthy adult female German blackheaded sheep with an average weight of 70.5 kg were included in the study. After intravenous induction (1 ml midazolam, 5 mg/kg propofol), anaesthesia was maintained with isoflurane delivered in 100% oxygen (1 l/min). Additionally, buprenorphine (10 mg/kg intramuscular), carprofen (4 mg/kg half intravenous, half subcutaneous), and fentanyl (0.005 mg/kg intravenous) were applied for analgesia. Systolic, diastolic, and mean blood pressure, electrocardiogram, rectal temperature, and haemoglobin oxygen saturation were monitored continuously during surgery.
All sheep were placed on their right side, exposing their left side for the surgical approach. Beginning from the dorsal aspect of the frontal extremity, a lateral incision of skin and skin muscle was performed following the leading edge of the latissimus dorsi muscle. After blunt penetration of soft fat layers, the latissimus dorsi muscle was undermined with exposure of the thoracodorsal vessel tree. Both the thoracodorsal artery and vein were prepared completely and released from the surrounding tissue over a distance of approximately 10 cm, leading to a straight arteriovenous trunk. Vessel branches were stripped-off using either electrocoagulation or vessel clips, depending on their diameter. Special attention was given to an uncompromised blood flow within the main arteriovenous trunk, whose continuity was left untouched. During preparation, pulse-synchronic movements of the main arteriovenous trunk were continuously present. Perfusion was additionally controlled using common microvascular surgery techniques, such as smoothing out of vessels against the direction of blood flow.
All animals in group A, group B, and group C were surgically supplied with an axial perfusion system. A cylindrical titanium cage of 20-mm diameter and 60 mm in length was bent intraoperatively over an individually created template with a central angulation of 30°. After implantation of the titanium cage into the latissimus dorsi region, the prepared arteriovenous trunk was placed into the central position of the titanium cage in order to provide axial perfusion of the prefabricated construct. Following the concept of modular construction, the internal structure was built with two β-tricalcium phosphate (β-TCP) cylinders of 14-mm diameter and 25-mm length (ChronOs; Synthes, West Chester, PA, USA) with a central passage of 7-mm diameter; these were modified preoperatively and sterilized again by gamma radiation. Intraoperatively, the cylinders were sliced lengthwise using piezoelectric surgery and were plugged onto the central arteriovenous trunk inside the titanium cage. According to the manufacturer’s data, the implanted β-TCP cylinders have an interconnective structure with a porosity of 60–80% and a pore size of 100–500 mm ( Fig. 1 A–D ).
Group A: in vivo application of autologous cancellous BMA
A total of 18 sheep were operated on following the experimental animal model protocol described above. Six of these sheep were assigned to group A. Using the whole organism as a natural bioreactor, the constructs of these animals were directly vitalized in vivo with osteogenic material from the iliac crest. Several bone biopsies (5 mm diameter) were harvested, morselized with an electric bone mill, and mixed with amorphous bone marrow, which was aspirated from the depth of the biopsy areas (BMA). The osteogenic material was placed around the β-TCP cylinders that had been sliced lengthwise and soaked in blood, until the titanium cage was completely filled, keeping the central arteriovenous trunk inside. Closure of the titanium cage was performed with two non-absorbable sutures.
Group B: in vivo application of nucleated cell concentrate (NCC)
Six sheep were assigned to group B. Using a bone marrow concentration system (MarrowStim; Biomet Biologics, Warsaw, IN, USA), the nucleated cell population could be extracted from the BMA, including mesenchymal stem cells (MSC). Compared to group A, this procedure reduced the peripheral blood dilution from the aspiration process. With a specially designed aspiration needle, a total of 60 ml anticoagulated BMA (54 ml BMA, 6 ml heparin solution with 1000 U/ml) was harvested from the iliac crest of each animal. After balance and spin in a centrifuge (3200 rpm, 15 min), the BMA was separated into three distinct layers: cell-poor plasma (CPP), NCC, and red blood cells (RBC). The NCC was suspended by vigorous shaking for 30 s and subsequently extracted from the concentration system with a sterile 10-ml syringe for construct vitalization.
Group C: isolation and in vitro cultivation of bone marrow stromal cells (BMSC) with scaffolds in a perfusion bioreactor system
A custom-made bioreactor system was used for cell seeding and culture of bone marrow mesenchymal cells. The system has been described and used for clinical case studies of tissue engineered bone previously. Briefly, it consists of a culture chamber hosting the porous scaffold and an inflow and two outflow valves. For the cell seeding, one outflow valve is closed and one is open to permit aspirates to propagate through the scaffold pores. The bioreactors were filled with saline containing 1000 U of heparin (Liquemin). The same solution was used to irrigate the aspiration syringe before its use. A stab incision was performed during a first operation at the most prominent aspect of the posterior iliac crest. A Jamshidi aspiration was inserted and pushed into the marrow cavity. The marrow was aspirated directly through a silicone tube into the porous β-TCP scaffold using an aspiration technique described previously. The bioreactors were rotated at 5° per min for 24 h. Then, the scaffolds were perfused with 500 ml of angiogenic medium containing DMEM Ham’s F-12 (Biochrom, Berlin, Germany), 10% foetal calf serum (FCS; Gibco, Darmstadt, Germany), 100 U of penicillin and 0.2 mg/ml streptomycin (Biochrom), 0.5 μg/ml amphotericin B (Biochrom), 5 μg/ml ascorbic acid (Sigma–Aldrich Chemie GmbH, Taufkirchen, Germany), 0.02 nM/ml dexamethasone (Merck, Darmstadt, Germany), and 30 ng/ml of a TLR-2/6 receptor agonist (MALP-2; Axxora GmbH, Lörrach, Germany), with concentrations recommended in a recent publication. The bioreactors were kept in an incubator at 37 °C and culture perfusion was installed with a media flow of 2 ml/min and 5% CO 2 . Every third day, 50 ml of culture medium was exchanged. After 21 days, the bioreactors were disconnected and transferred to the operating room at a temperature of 5 °C. After removal from the bioreactors, the seeded scaffolds were used for transplant construction as described above and transplanted into the latissimus dorsi region of six sheep.
To control the number of progenitor cells aspirated and transplanted on the scaffolds, aspirates were analyzed using a cell flow cytometer (SE-5000 Automated Hematology Analyzer; Sysmex, Kobe, Japan). The same amount of aspirate that was kept in the bioreactor was used for a density centrifugation (Bicoll; Biochrom, Berlin, Germany) and plated on a 125-cm 2 culture flask (Nunc, Fisher Scientific, Loughborough, UK). The same media and culture conditions were used for these samples. The number of fibroblast colony forming units (F-CFU) was assessed after 21 days.
3D imaging, clinical and histological evaluation
Conventional computed tomography (CT) examinations were performed directly after surgery, 3 months after surgery, and 6 months after surgery at the time of sacrifice, to determine construct position and ceramic resorption. All animals were killed after deep sedation (midazolam 1 mg/kg intramuscular, propofol 5 mg/kg intravenous, pentobarbital 80 mg/kg intravenous). After explantation from the latissimus dorsi muscle, the cylindrical titanium cage was removed from the constructs, which were manually evaluated and divided into three parts of equal length on a saw bench. After fixation in 3.5% neutral buffered formalin and embedding in methylmethacrylate (MMA), the central part with the 30° angulation was sectioned lengthwise according to the axis of the constructs with a modified inner-hole diamond saw for evaluation of the junction between the two cylinders inside the constructs. The lateral straight arms of the constructs were sectioned perpendicular to the axis of the cylinders inside the constructs. Undecalcified slices of 30 μm thickness were surface-stained with alizarin–methylene blue for standard light microscopy and histomorphometric analysis. Five equally distributed slices of every specimen were generated. Digital images were obtained using a Zeiss AxioImager MI microscope fitted with an AxioCam MRc digital camera and AxioVision 4.5 software (Carl Zeiss, Oberkochen, Germany). The AxioVision module MosaiX was used to reassemble digital images of the entire cross-section of the constructs, which provided the basis for further analysis. Total bone area, residual ceramic area, and fibrous tissue area were quantified using the image analysis software analySIS 3.01 (Olympus Soft Imaging Solutions, Münster, Germany).