Repair for large bone defects due to fractures and cancers is a huge challenge as these defects do not spontaneously heal and require additional treatments such as bone graft and/or implants to enhance new bone regeneration. Many kinds of bone grafts have been used to repair large bone defects such as autograft, allograft, and artificial bones. Autograft is regarded as the excellent scaffold for bone regeneration. However, there are several problems in autograft, including morbidity of the donor site, insufficient amount of bone graft that can be harvested, and additional surgery is required. Allograft is another choice for bone graft, but it could induce immunologic reaction in the recipient and the risk of transferring diseases.

Recently, a large number of artificial bones with various compositions are commercially available. Calcium phosphate materials, such as hydroxyapatite (HA) and tricalcium phosphate (TCP), have been widely used as artificial bones in orthopedic, maxillofacial, and plastic surgeries due to their high biocompatibility and osteoconductivity. However, HA has a low degradation rate and takes a longer period to be replaced by new bone tissues in vivo. On the contrary, TCP has been widely used since it is replaced with host bone tissues.

TCP is subdivided into alpha and beta form by its crystallinity. The granule form of β-TCP is commercially available, but it is irregular in shape and size and is fragile in mechanical strength, therefore its use is restricted for non-load-bearing sites. α-TCP has larger mechanical strength with the surface of larger crystals. α-TCP is a gradually degradable osteoconductive material and is suggested to be a better choice in TCP artificial bones. Currently octacalcium phosphate (OCP) has become an important artificial bone, and is supposed to be a precursor of biological apatite such as bone and tooth. OCP could stimulate osteoblastic cell differentiation in vitro, and exhibited biodegradability, osteoconductivity, and osteoinductivity in vivo. In addition, it was reported that the cement comprised of OCP and α-TCP had higher compressive strength.

Another factor influencing on new bone formation and mechanical strength is the macrostructure design. The artificial bone with adequate connective interpores was expected as an ideal bone graft, because these interpores may provide the space for better bone ingrowth capability. Bone growth within granules must increase the mechanical strength with new bone regeneration in these interconnected pores.

Our group developed a novel tetrapod-shaped granular artificial bone (TetraboneR) made from a mixture of α -TCP and OCP with the size of 1mm. Both rupture strength of a single particle and elastic modulus of aggregated particles of TetraboneR were significantly higher than those of β-TCP granules in vitro. Therefore, the purpose of this thesis was to clarify the new bone formation and mechanical properties of TetraboneR when repairing the femoral defect in animal models.

In chapter 1, I investigated the long-term effect of implantation of TetraboneR into the femoral condyle defect in rabbits on new bone regeneration and its safety, and compared with those by β-TCP granules. Eighteen male New Zealand white rabbits were used. The bone defect (5mm in diameter and 8mm in depth) was made at both lateral femoral condyles. Then, the defects were filled with TetraboneR (N=15) or β-TCP granules (OsferionR) (N=15) or left empty (control: N=6). Rabbits were euthanized at 4,13, and 26 weeks after surgery. After euthanasia, the condyles were collected and examined on gross observation, micro-CT, and histology. There were no clinical side effects in any rabbits receiving TetraboneR during 26 weeks of the experimental period. However, granule leakage was observed in the β-TCP granule group at 4 weeks of implantation.

TetraboneR was retained well inside the defect with less granular absorption at 26 weeks of implantation. The opening of the defect was covered with connective tissues and concave-shaped in the β-TCP and control groups. β-TCP granules were rapidly resorbed at 4 weeks, which might lead the concave shape at the opening of the defect due to the fragile property of the granule at 13 and 26 weeks of implantation. In addition, new bone area in the TetraboneR group was more than that in the β-TCP granule group at 13 and 26 weeks. These results suggested that β-TCP granules could not maintain the shape of the defect due to its rapid resorption, especially in the central area of the defect without new bone formation. It was indicated that TetraboneR had better osteoconductivity than β-TCP granules in the long-term implantation.

In chapter 2, I investigated the mechanical properties of TetraboneR implanted into the femoral defect of canine cadavers and compare with the β-TCP granules. Fourteen beagle dog cadavers were used and rectangular trabecular bone blocks with the shape of 14mm X 14mm X 8mm and the defect of a 10 mm cylindrical hole in the center were made from each distal femoral condyle. The defects were filled with artificial bones and divided into 3 groups; TetraboneR group (N=8), β-TCP granule group (OsferionR) (N=8), and control group (PEG gel) (N=8). Additionally, 4 other femoral bone blocks without the cylindrical hole were used as an intact control.

The femoral bone blocks were mounted on InstronR mechanical test system (Instron-3365) to determine the load-deformation changes of the specimens in vertical compression until fracture, and the ultimate compressive load and elastic modulus were calculated from the mechanical test recordings. The ultimate compressive load of the TetraboneR group was almost half of the intact bone and was significantly higher than those of the β-TCP granule and control groups. The elastic modulus was significantly higher in the TetraboneR group than those of the β-TCP granule and control groups.

In the defect insertion testing, 15 femoral condyles with a tunnel defect (10 mm in diameter) were made and divided into the same 3 groups. The specimens were fixed on a rheometer with bone cement, and the rod (5 mm in diameter) was vertically inserted into the exposed surface of the graft material (10mm in diameter). The force-displacement changes were measured and the slope of the initial linear portion of the force-displacement curve was defined as compressive stiffness. The slope of force-displacement curve was higher in the TetraboneR group than other groups, and the compressive stiffness was significantly higher in the TetraboneR group than other groups. In conclusion, it was confirmed that TetraboneR implanted into the canine femoral defect model showed better mechanical properties than β-TCP granules in vitro.

In chapter 3, I investigated the effect of TetraboneR implanted into the femoral defect in dogs by evaluation of new bone formation and mechanical strength. Seven male beagle dogs were used and a tunnel defect (10 mm in diameter) was made at the both femoral condyles. The tunnel defects were filled with TetraboneR (N=5), β-TCP granules (N=5), or without filling (control group) (N=4). All dogs were euthanized at 8 weeks after surgery.

No clinical side effects were observed in all experimental dogs. On gross findings, the opening of the defect was concave in the β-TCP granule and control groups, while smooth in the TetraboneR group. On radiography, CT, and micro-CT analysis, the center of the defect in the β-TCP granule group was shown as dark appearance, suggesting the earlier resorption of the β-TCP granules and no new bone formation, whereas TetraboneR was retained in the defect at 8 weeks of implantation. Compressive stiffness of the defect in the TetraboneR group was maintained almost 80% of the intact bone and was significantly higher than that of the β-TCP granule group. These results suggested the better mechanical strength of the defect implanted with TetraboneR than with β-TCP granules after 8 weeks of implantation.

On histology, new bone distribution was significantly higher in the TetraboneR group than that in the β-TCP granule group, though new bone area was similar in both groups. The new bone tissues in the TetraboneR group were fully distributed in the defect area, while in theβ-TCP granule group, granules were resorbed and the new bone tissues were partially distributed around the defect margin. These results indicated that the interconnectivity of the intergranular pores were effective for new bone invasion in the TetraboneR group. In addition, TetraboneR had better osteoconductivity than β-TCP granules.

In conclusion, although TetraboneR was resorbed at a much slower rate than β-TCP granules, TetraboneR provided much higher mechanical strength and better osteoconductivity in the defect than β-TCP granules. These results encourage the clinical application of TetraboneR for the large bone defects at load-bearing sites in the clinical practice.

腫瘍摘出などに伴う巨大な骨欠損に対しては、骨移植が必須の治療法である。従来、これらの症例に対し、自家骨あるいは同種骨の移植が用いられてきたが、自家骨移植には得られる骨量の制約、骨採取のための手術の必要性などの問題がある。また同種骨については、日本にそれを供給するシステムがないこと、あるいは拒絶反応の可能性や感染の伝搬というリスクも存在する。

このような背景をもとに、多くのセラミック系人工骨が開発されてきた。これらの中で骨組織に類似するハイドロキシアパタイト、リン酸三カルシウム(TCP)などが広く用いられてきた。TCPにはα型とβ型の二つのタイプがあり、何れも徐々に体内で自家骨に置換される。既にβTCPは市販されており、その有用性も確認されている。一方、リン酸オクタカルシウム(OCP)が注目されており、その強度ならびに骨誘導能はTCPより高いといわれている。

著者らのグループは、αTCPとOCP(表層部分)から成る約1mmのテトラポッド型の顆粒状人工骨(TB)を開発した。これは市販のβTCP 顆粒に比較して強度が高く、かつ顆粒同士を噛み合わせたときにできる顆粒間の連通孔が、血管新生に好ましい足場を提供する、といいう仮説で設計されたものである。

本研究の目的は、TBの骨形成能ならびに力学的特性に関する有用性を、市販されているβTCP顆粒を対照として、動物モデルならびに死体を用いて評価することにある。

第1章では、ウサギの大腿骨遠位骨顆部に円筒形欠損(直径5mm、深さ8mm)を作成し、TBならびにβTCP顆粒を移植し、非移植群を対照としてその骨形成能と臨床的有用性を検討した。ウサギは移植後4、13、26週後に安楽死し、μCTおよび組織学的に評価した。その結果、13および26週において、欠損部に対する新生骨面積はTB群で有意に多く、欠損部全体に新骨組織が認められた。βTCP顆粒は移植後早期に吸収され始め、13、26週ではほとんど完全に吸収されていた。そのため、欠損部の表面はくぼんでおり、欠損部の骨脆弱性が確認された。また、26週という長期にわたり、TBに関する副作用は認められなかった。

第2章では、成犬死体の大腿骨遠位骨顆に直径10mmの貫通欠損を作成し、TBおよびβTCP顆粒を移植したときの力学的強度を測定した。まず移植部を含む骨ブロック(14x14x8mm)に対し、垂直荷重をかけたときの最大圧縮荷重と弾性率を測定した。最大圧縮荷重について、TB群は欠損のないintact骨の約半分の強度を有しており、大きな荷重にも耐えうるのではないかと考えられた。またこの値は、有意にβTCP 群より高い値であり、弾性率も同様に有意に高値を示した。次に、大腿骨そのものを固定し、欠損部にバーを刺入して測定する欠損刺入試験を行った。その結果、荷重ー変位線図でもTB群がβTCP群より有意に高く、かつ欠損部そのものの圧縮強度も有意に高いことが示された。以上から、TBの力学的強度はβTCP よりも強く、大きな骨欠損部に移植した場合でも、より高い荷重にも耐え得ることが示された。

第3章では、ビーグル成犬を用い、同様の骨欠損を作成してこれらの人工骨を移植し、その有用性を評価した。まず臨床経過を画像診断により評価するとともに、8週後に安楽死した後、組織学的な骨形成評価ならびに力学的強度を測定した。

その結果、TB群では顆粒は安定して欠損部に存在したが、βTCP 群では、顆粒の吸収に伴って、欠損部中央の骨密度は低下した。この群の安楽死後の肉眼所見では、欠損部の表面は陥没しており、早期のβTCP 吸収に伴う力学的強度の低下が示唆された。移植8週後の力学的評価では、TB群の圧縮強度はintact骨の約80%の強度を有していたのに対し、βTCP群では、有意に低値を示した。

移植8週目の組織では、新生骨量は両群に有意差はなかったものの、新生骨の分布が大きく異なっており、TB群では欠損部に広く分布しているのに対し、βTCP群では、欠損部の周囲に分布していた。このことは、早期に吸収されたβTCP 顆粒のため、欠損部中央に血管新生が進展せず、そのために中央部の骨形成が不十分であったものと推察された。一方、TB群では顆粒間の連通孔内に血管新生が生じ、それに伴って骨形成が進展したものと考えられた。また、連通孔に形成された骨組織も、欠損部の強度維持に役立っているものと推察された。

以上要するに、本研究は新たに開発されたテトラポッド型顆粒状人工骨の巨大骨欠損修復時における骨形成能ならびに力学的有用性を示したものであり、学術上、臨床上その貢献するところは少なくない。よって審査委員一同は本論文が博士(獣医学)の学位論文として価値あるものと認めた。