When bone morphogenetic protein (BMP) is delivered to matrices in vivo may affect tissue engineered bone constructs for jaw reconstruction after cancer surgery. This study compared the effects of BMP application at different times after matrix implantation for heterotopic bone induction in a rat model. Hydroxyapatite blocks were implanted unilaterally onto the surface of the latissimus dorsi muscle. A second block was implanted onto the contralateral muscle after 1, 2 or 4 weeks and 200 μg rhBMP-2 was injected into the blocks on both sides. Bone formation and density inside the blocks was analysed by CT and histology. 8 weeks after BMP application increases in bone density within the scaffolds were most pronounced in the simultaneous application group (179 HU). Less pronounced increases were observed for the 1 (65 HU), 2 (58 HU) and 4 (31 HU; p < 0.0001) week delay group. Homogeneous bone induction started from the central channel of the blocks. Capillaries and larger vessels were seen in all constructs, samples receiving delayed BMP treatment demonstrated significantly greater neovascularization. Delayed application of BMP was less effective for heterotopic bone formation than simultaneous application. A central channel allows homogeneous bone induction directly from the centre of the blocks.
The in vivo tissue engineering of heterotopic bone by intramuscular endocultivation, where the patients serve as their own bioreactor, has yielded customized vascularized bone grafts that have already been used in humans for jaw reconstruction after ablative tumour surgery. Midface reconstructions are difficult because of the complex anatomy. Three dimensional (3D) planned reconstructions are being studied based on computed tomography (CT) or magnetic resonance imaging (MRI) data. The origin of materials used ranges from hydroxyapatite or derivates based on polylactic acid (PLA) or polylactic-co-glycolic acid (PLGA).
This cultivated bone is mostly cancellous with no cortical layer. Excess bone formation is often observed outside the matrices. It remains a challenge to achieve thorough and homogeneous bone induction throughout the entire graft. In a pilot study, the authors administered BMP-2 for bone induction 4 weeks after the implantation of the scaffold (delayed application). They used bovine hydroxyapatite (HAP) blocks but did not observe any improvement in bone induction with a delayed application. One reason was that excessive soft tissue ingrowth may have prevented the penetration of the BMP deep within the scaffolds. Earlier studies tested the use of a PLA membrane to avoid soft tissue ingrowth into HAP scaffolds placed onto the mandible. This technique enhanced homogenous bone formation.
The results of the first study led the authors to modify the matrices and examine the effect of delays shorter than 4 weeks (this study). The special design of the matrices included a central channel to allow BMP injection directly into the centre of the blocks and a porous structure.
The novel aspect of this study was the administration of cytokines for bone induction and cell differentiation at several different time intervals. Usually, the scaffold and cytokines are implanted simultaneously at the same operation in animal models. At the time of implantation no vascularization or osteoblasts are present in the centre of the scaffolds. The cellular machinery for bone formation has to migrate from the surface of the scaffold to the centre. The authors reasoned that a delay in the application of cytokines might allow undifferentiated mesenchymal cells to penetrate into the scaffold, thereby enhancing the number of stem cells within the scaffold at the time of application of the cytokines (BMP). The authors reasoned that should this occur, vascularization within the scaffold would be promoted and there would be an increased likelihood of homogenous osteogenesis within the scaffold, with consequently higher bone density.
The aim of this study was to evaluate heterotopic bone formation on a novel HAP scaffold produced using computer aided design (CAD) in a rat model. In particular, the influence of soft tissue ingrowth with delayed application of BMP-2 at different intervals was compared to the simultaneous placement of the matrix together with the BMP-2.
Materials and methods
24 female Lewis rats (4 months old, about 200 g each) were used for this study. The study protocol was in accordance with the German Animal Welfare Act (Animal Experiment Permit V-742-72241.121-14 (39-5/04)). All animals received food and water ad libitum. Every surgical procedure and each CT examination was performed under general anaesthesia with ketamine (10%, 1 μl/g body weight, intraperitoneal) and xylazine (2%, 0.6 μl/g body weight). The implanted HAP blocks were specially designed and produced as previously published. A spray-dried granulate HA19 (BioCer Entwicklungs-GmbH, Bayreuth, Germany) was chosen as the raw material for 3D printing of HAP scaffolds.
Each animal was implanted with a HAP block on the left side. After different delay periods allowing for soft tissue ingrowth in the first block, a second block was inserted on the right side and BMP was applied to both blocks ( Fig. 1 ). This allowed the authors to examine simultaneous and delayed application of BMP and its effect on heterotopic bone formation in vivo in the same animal. The animals were separated into three groups (8 animals each): the first group had a delay period of 1 week, the second group a period of 2 weeks and the third group a period of 4 weeks. These groups are referred to as the ‘1 week delay’, ‘2 week delay’ and ‘4 week delay’ groups. The delay period reflects the time given to allow for soft tissue ingrowth into the first block. The right side (simultaneous application) served as a control, where no time for soft tissue ingrowth was given.
The surgical procedure and the amount of rhBMP-2 solution applied are reported in earlier publications. The skin was shaved and disinfected before surgery. Over a median incision on the thoracic spine, the latissimus dorsi muscle on the left side was displayed and a pouch above the muscle was created. A special CAD HAP block (about 2 cm × 1 cm × 1 cm) with one side rounded off and a central channel ( Fig. 2 ) was implanted. The pouch and the skin were sutured separately.
In the second operation, a similar HAP block was inserted through the same approach on the right side. Additionally, 1 ml of rhBMP (containing 200 μg BMP-2; Wyeth Pharma GmbH, Münster, Germany) was applied with a syringe directly into the central channel of each block on both sides, giving 400 μg BMP-2 per animal. The block was porous and the BMP-2 penetration into surrounding soft tissue was avoided by surgical creation of a pouch.
CT examinations under general anaesthesia were performed (Somatom, Siemens AG; 120 kV, 210 mA, 46.82 mGy) to follow increases in bone density until 8 weeks after the second operation. Blinded examiners performed bone density evaluation.
Fluorochrome labelling by intraperitoneal injection started 1 week after the second operation and the BMP application as follows: xylenol-orange (6% in NaHCO 3 , 1.5 ml/kg body weight; Sigma–Aldrich, Steinheim, Germany) was applied at weeks 1 and 5 after the second operation; calcein-green (1% in 2% NaHCO 3 , 5 ml/kg body weight; Sigma–Aldrich, Steinheim, Germany) was applied at weeks 2 and 6; alizarin (3% in 2% NaHCO 3 , 0.8 ml/kg body weight; Sigma–Aldrich, Steinheim, Germany) was applied at week 3; tetracycline (1 mg/kg body weight, Doxycyclin Ratiopharm SF, Ratiopharm, Ulm, Germany) was applied at week 4.
At the end of the observation period (8 weeks after the second operation), blocks on both sides were excised together with the surrounding soft tissue (this technique is described in detail in an earlier publication ). Specimens were dehydrated in a graded series of alcohol solutions and embedded in wax. The specimens were ground to a width of 80 μm for microradiography, 40 μm for fluorescence microscopy (Mikrophot-FXA, Nikon, Germany) and to 8 μm for staining (Masson-Goldner, toluidine) and light microscopy. Microradiography was performed as previously described. Haematoxylin–eosin (H–E) staining was performed on 4 μm sections, which were analysed semi-quantitatively for vasculature. A defined area within the construct was analysed. Samples were scored at 40× magnification; those with construct areas <70 mm 2 were scored in total whereas those with an area >70 mm 2 were partially scored ( i.e. every 4 fields of view at 40× magnification). The scoring method was adapted to assess four sizes of blood vessels <10 μm, 10–30 μm, >30 μm or longitudinal vessels.
Bone density was measured in Hounsfield Units (HU) from CT scans. Data were analysed by ANOVA with the two factors: group (0, simultaneous application; 1, delayed by 1 week; 2, delayed by 2 weeks; 4, delayed by 4 weeks) and time (0 and 8 weeks after BMP application). For further analysis, multiple comparisons after Scheffé were calculated. For expected values, the corresponding 95% confidence intervals were computed.
Vascularization data are presented as the mean number of blood vessels per mm 2 on the tissue section (±SEM). Differences between the means of the two treatment groups (delayed vs. no delay in BMP treatment) were tested for statistical significance. Normally distributed data was tested using a 2-sample t -test and all non-normally distributed data tested using the non-parametric Mann–Whitney U -test.
Bone density evaluation
The increases in bone density were higher in the control group (179 HU; 95% CI: 154–205 HU; Fig. 3 ) where application of BMP occurred simultaneously with implantation of the HAP blocks. Lower rates were observed for the 1 week (65 HU; 95% CI: 39–91 HU) and 2 week (58 HU; 95% CI: 32–84 HU) delay groups. The lowest increase was seen in the 4 week delay group (31 HU; 95% CI: 5–57 HU). The group effect in general was statistically significant ( p < 0.0001). For the following combinations, the p -values of the group effect were below 0.05: 0 and 4, 1 and 2, 1 and 4, 2 and 1, 2 and 4 and 4 against all others ( Table 1 ).
|Simultaneous application||1 week delay||0.076|
|Simultaneous application||2 week delay||0.681|
|Simultaneous application||4 week delay||0.000|
|1 week delay||Simultaneous application||0.076|
|1 week delay||2 week delay||0.004|
|1 week delay||4 week delay||0.023|
|2 week delay||Simultaneous application||0.681|
|2 week delay||1 week delay||0.004|
|2 week delay||4 week delay||0.000|
|4 week delay||Simultaneous application||0.000|
|4 week delay||1 week delay||0.023|
|4 week delay||2 week delay||0.000|
The statistical evaluation of different effects revealed that no individual animal-specific effects accounted for the increase; only group-dependent effects accounted for the change in density. Linear F -tests indicated the absence of systematic curves in all groups, indicating that a linear interpretation for bone density increase was sufficient.
No animal developed evidence of infection, seroma or haematoma formation over the implanted HAP blocks. Two animals died in each group during general anaesthesia. This occurred prior to the planned end of the observation period. They were therefore excluded from the evaluation without replacement.
Microradiography and fluorescence microscopy
Microradiography indicated homogenous new heterotopic bone formation occurred in the control group and in the 1 and 2 week delay groups. All delayed groups demonstrated bone induction commencing from the central channel. The 4 week delay group showed the least evidence of bone induction ( Fig. 4 ).
Toluidine staining revealed new bone formation in each group ( Fig. 5 ). The integration of fluorochromes revealed evidence for new bone formation especially during weeks 2 and 3 (green bands for calcein green and red bands for alizarin) of each group. Bands for all fluorochromes were found, showing ongoing bone formation afterwards ( Fig. 6 ) and for the entire period.