2.3. Bone morphogenetic protein

2.3.1. TGF-β superfamily

BMPs belong to a group of proteins called TGF-β superfamily, and this gene family currently includes at least 43 members (Wozney & Rosen 1998). The proteins of the TGF-β superfamily regulate many different biological processes, including cell growth, differentiation and embryonic pattern formation (Zhu et al. 1999). This group of proteins includes, among others, transforming growth factors (TGF-β 1 through TGF-β 5), BMPs and growth and differentation factors (GDFs) (Burt & Law 1994). BMP1 is the only BMP that is not a member of the TGF-β superfamily, but is a procollagen C-proteinase, which is the prototype of a family of putative proteases implicated in developmental pattern formation in diverse organisms (Kessler et al. 1996, Li et al. 1996). BMPs 2–16 are the presently known members of the BMP superfamily (Dube & Celeste 1996a,b, Wozney & Rosen 1998), and they can be divided into different subgroups according to how closely they are related to each other structurally (Table 1). For example, BMP-2 and BMP-4 are 92 % identical, while BMP-5, BMP-6 and BMP-7 are 90 % identical (Wozney 1992).

Bone morphogenetic proteins are dimeric molecules with two chains held together by one disulphide bond. Each monomer consists of about 120 amino acids with seven canonical cysteine residues (Reddi 1998a).

BMPs 2, 4 ,5 ,6 and 7 have been shown to be fundamentally important regulators of skeletal tissue formation and repair (Wozney et al. 1990, Cook et al. 1994a, Riley et al. 1996, Wozney & Rosen 1998, Cook 999). Different BMPs are not identical in their osteoinductive potential. For example, BMP 5 is needed in larger amounts to induce the same amount of bone compared to BMP 2 or 7 (Wozney & Rosen 1998).

2.3.2. Extracted BMPs

After Urist’s pioneering experiments, BMP was extracted from many different species, including rabbit (Urist et al. 1979), pig (Wu & Hu 1988), cow (Wang et al. 1988), dog (Heckman et al. 1991), baboon (Ripamonti et al. 1992), reindeer (Jortikka et al. 1993b), moose (Viljanen et al. 1996) and human (Urist et al. 1983).

The separation of BMP is extremely difficult, because it is almost totally insoluble in conventional buffer solutions. Furthermore, when extracted from demineralized bone matrix, it appears as high-molecular-weight protein aggregates. To break down these aggregates, a number of sequential precipitation-solubilization steps are required (Marttinen et al. 1992, Jortikka 1993a). Briefly, the extraction process involves the following steps:

1. mechanical stripping of long diaphyseal bones

2. pulverization of bone material

3. demineralization in HCl

4. extraction by GuHCl or urea

5. ultrafiltration

6. chromatography

Native BMP is present in cortical bone in minute amounts, approximately 1–2 µg BMP/kg of cortical bone. Thus, large amounts of bone are needed to produce sufficient amounts of BMP for experiments.

Native bovine BMP is the most frequently used native BMP in animal studies because of the availability of bovine bone and the proven effect of bovine BMP. Bovine BMP has a molecular weight of about 18 KD (Urist et al. 1982, Bessho et al. 1989, 1991a, 1991b, Shibahara et al. 1995). Bovine BMP has also been the origin of recombinant BMPs, since the amino acid sequence has been derived from a highly purified preparation of BMP from bovine bone (Wozney et al. 1988).

2.3.3. Recombinant BMPs

With the purification of human osteogenic proteins of sufficient purity to provide amino acid sequence data, complementary DNA clones were isolated, cloned and expressed in host cells. Thus, the human BMPs 1 through 7 were found (Wozney et al. 1988, Wozney 1989, Celeste et al. 1990). The recombinant BMPs 2, 4 and 7 have been shown to induce bone in many experiments and are now also being tested in clinical studies (Boden 1999).

Although it has been shown that a single rhBMP is able to induce bone formation ectopically (Wang et al. 1988), it is interesting that the amount of human rhBMP necessary to produce bone induction in vivo is more than 10 times higher than that of highly purified bovine extracted BMP. Recently, Bessho et al. (1999) demonstrated this difference in effect between purified human BMP derived from human bone matrix and recombinant human BMP. This fact suggests that native BMP activity is a combination of the activities of different BMPs or represents synergistic activity between them (Wang et al. 1990).

2.3.4. Functions of BMPs

The hallmark of bone morphogenetic protein activity in vivo is the induction of new bone. The standard method for assaying BMP is its intramuscular implantation into a mouse or rat and the estimation of new bone induction by radiology and histology (Urist & Strates 1971). In rat bioassay, other growth factors and extracts from other connective tissue matrices prepared according to the BMP extraction procedure do not have osteogenic activity, which means that BMPs are the only growth factors with a known ability to stimulate the differentiation of mesenchymal stem cells in the chondro- and osteoblastic direction (Reddi et al. 1987, Aldinger 1991, Chen et al. 1991, Solheim 1998). BMPs initiate, promote and maintain chondrogenesis and osteogenesis, and BMPs also have many extraskeletal functions, as they regulate the development of several embryonic structures, including the kidney, lung and gut (Hogan 1996, Reddi 1998b). It seems that, in a mature animal, bone repair after injury is similar to bone formation in an embryo, suggesting analogous mechanisms for the control of bone formation in adult and embryonic skeletons (Rosen & Thies 1992).

Bone morphogenetic proteins exert their effects through receptors, which are members of a larger family of serine threonine kinases, including the receptors for transforming growth factor betas, activins and inhibins (Massagae et al. 1994). BMP receptors are of two types, type I and type II. These receptors phosphorylate cellular Smad proteins, which transcriptionally activate target genes (Dewulf et al. 1995, Reddi 1998a, Laitinen 1999, Miyazono 1999).

The interactions of BMPs with other agents remain quite obscure. There is some evidence that prostaglandin E1, for example, has promotive effects on the osteogenic activity of rhBMP (Ono et al. 1996).

When an osteogenic implant is implanted, it activates a series of cellular events, including chemotaxis of pluripotential mesenchymal cells into the implant site, differentiation of these cells into chondroblasts and osteoblasts, removal of calcified cartilage, and population of new bone with bone marrow elements. The bone morphogenetic proteins induce new bone formation through endochondral ossification, where cartilage forms first and is subsequently replaced by bone (Sampath & Reddi 1981, Wozney & Rosen 1998).

2.3.5. BMP carriers

To enhance osteoinduction, bone morphogenetic proteins must be mixed with an appropriate carrier substance, since the proteins are soluble within biologic fluids. Although there is no absolute need for a delivery system, if a sufficient amount of bone morphogenetic protein is applied, bone formation can be observed (Forslund & Aspenberg 1998, Wozney & Rosen 1998), a carrier system is required to optimize the osteogenic activity of BMP (Lindholm & Gao 1993, Ripamonti 1993). It has been shown that the carrier material may have an effect on the pharmacokinetics of BMP on the basis of different release patterns (de Groot 1998, Winn et al. 1999).

Overall, the development of appropriate osteoconductive carriers has not progressed as rapidly as the isolation and synthesis of growth factors. This has significantly slowed down the development of clinically successful biosynthetic composite implants (Lane et al. 1999b).

Theoretically, the carrier material will have to meet the following requirements (Aldinger et al. 1991):

1. relative insolubility in physiological conditions

2. biodegradability

3. protection against proteolytic activities

4. substrate for cell adhesion and proliferation

5. immunologically inert

6. slow release of BMP through controlled biological degradation

7. mechanical stability in bridging bone defect

Many different carrier materials have been used in a variety of animal models, in which bone morphogenetic proteins have been tested (Cook et al. 1994a, Hollinger & Seyfer 1994, Wozney & Rosen 1998, Winn et al. 1999), but the optimal carrier material for BMPs still remains to be found. The optimal type of carrier material used will probably depend on the clinical indication to which the morphogenetic protein will be applied (Wozney & Rosen 1998).

The carrier material can be in the form of blocks, granules, paste, solution or as a self-setting cement (Kamegai et al. 1994, Ohura et al. 1999).

Carrier materials can be classified based on different criteria, such as inorganic versus organic, biological versus non-biological and bioedgradable versus non-biodegradable (Viljanen 1997)

Broadly speaking, the carrier materials for BMP can be divided into five major categories:

1. Demineralized bone matrix

2. Collagenous materials

3. Resorbable synthetic polymers

4. Calcium phosphate materials

5. Others

Demineralized bone matrix (DBM) has been used in many studies as a carrier material for BMPs, extracted with GuHCl to remove endogenous bone inductive activity (Cook et al. 1994a,b, Sciadini et al. 1997a), or as a commercially available preparation of demineralized freeze-dried human bone powders (Niederwanger & Urist 1996). The experiments of Cook et al. (1994a, 1994b) and Sciadini et al. (1997a) demonstrated the suitability of demineralized bone matrix as a carrier for BMP in an animal long bone defect model. Toriumi et al. (1993) successfully repaired a 3 cm full-thickness mandibular defect in a dog with allogeneic DBM mixed with recombinant BMP-2. In one study, where different carrier materials for rhBMP-2 were compared in canine periodontal defects, DBM and Bio-Oss (sintered bovine bone) performed well compared to collagen, PLA and PGA, although the authors concluded that other impediments to their clinical use still exist (Sigurdsson et al. 1996). Immunogenicity remains a problem in demineralized bone matrix.

A tentative way to solve this problem, namely autolyzed, antigen-free, allogeneic bone (AAA), was developed by Urist and co-workers. AAA bone was later used in clinical studies by Johnson et al. (1990, 2000) with promising results. AAA cortical bone has undergone antigen extraction without significant alteration of the residual structural integrity of the cortical graft (Johnson et al. 1990).

Collagenous materials are superior in compatibility, because collagen is the major protein component of hard and soft tissues. A range of collagenous materials have been used in different studies, including collagen sponges and pastes (Sampath & Reddi 1981, Takaoka et al. 1988, Bessho et al. 1991a, Takaoka et al. 1991, Gao & Lindholm 1993a, Lindholm et al. 1992, Cook et al. 1995). However, immunogenicity and inferior osteoconduction limit the suitability of this material as an ideal carrier for BMP. The telopeptides of type I collagen are thought to be responsible for causing an immunogenic response when introduced into xenogeneic hosts. To eliminate this problem, Takaoka et al. (1991) used filtration to remove telopeptides. Telopeptide-depleted collagen as a carrier for BMP was found to be superior to conventional collagens in ectopic bone formation.

In the recent years, possibly the greatest interest has focused on resorbable synthetic polymers, such as polylactide (PLA) and polyglycolide (PGA), which are members of a large family of poly-alpha-hydroxy-acids. Polylactide is a synthetic thermoplastic polymer of cyclic diesters of lactic acid. Polylactic acid has two optically active stereoisomers, poly-L-lactic acid (PLLA) and poly-D-lactic (PDLA) (Tielinen 2000). The physical properties of the copolymers of L-lactic acid and D-lactic acid (PDLLA) are dependent on the relative amounts of L- and D-monomers. Their advantages include the synthetic nature of the system and the accumulated clinical and regulatory experience of PLA (Wozney & Rosen 1998).

Heckman et al. (1991) treated canine radial defects with BMP using both demineralized bone matrix (DBM) and polylactide carriers. The former had no effect on the healing of the defect, but the polylactide-BMP composite led to union in all cases. Some investigators have found that polylactic and polyglycolic acid porous microspheres, when combined with an appropriate dose of rhBMP-2, appear to be equally effective as inactivated demineralized bone matrix (Kenley et al. 1994, Muschler et al. 1994). Boström et al. (1996) used rhBMP-2 with a paste-like polylactide, treating rabbit ulnar defects with success. Hollinger and Leong (1996) suggested that poly-alpha-hydroxy acids are suitable carriers for BMPs based on some preclinical studies. rhBMP-2 was able to heal large segmental defects in sheep, when used with a PDLLA carrier in sheep femur (Kirker-Head et al. 1998). Zegzula et al. (1997) demonstrated the suitability of PDLLA as a carrier material for rhBMP-2 in rabbit radial diaphyses. In dentistry, synthetic polymers have also proven to be useful as carriers for BMP (Saitoh et al. 1994, Alpaslan et al. 1996). In an attempt to quantify osteoinductivity, Winn et al. used a measure of radioactivity to quantify rhBMP-2 pharmacokinetics, radiomorphometry, histomorphometry and alkaline phosphatase activity. The results showed that deorganified bovine bone resulted in an initial burst release of morphogen, but thereafter appeared to bind irreversibly a fraction of rhBMP-2. Collagen and PDLLA carriers showed a sustained release, and the latter also a dose-dependent release pattern (Winn et al. 1999).

Calcium phosphate materials, including coralline, hydroxyapatite, tricalcium phosphate and their composites, have been proposed as potential carrier materials for BMP. They resemble bone tissue structurally and are usually biocompatible, but their variable and often extremely slow biodegradation makes them suboptimal as carriers (Lane et al. 1999a).

Hydroxyapatite (HA) is a material that has been used widely in animal studies as a carrier material for various BMPs. It has been used in ectopic muscle implantation, in a skull defect model, under the periosteum of parietal bone and in mandibular bone defects, and a combination HA-BMP proved to be more effective than HA alone in all these studies (Takaoka et al. 1988, Damien et al. 1990, Horisaka et al. 1991, Ono et al. 1995, Asahina et al. 1997, Koempel et al. 1998). The effect of a HA-BMP combination in spinal fusion was demonstrated by Boden et al. (1999). The addition of collagen or bone marrow has further enhanced the osteogenic potential of the HA-BMP composite (Yoshida 1999, Noshi 2000). It has been suggested that the geometrical configuration of hydroxyapatite may be an important factor in osteogenesis (Magan & Ripamonti 1996, Kuboki et al. 1998).

Natural coral has been used in animal bone defect models with good results (Gao et al. 1997, Sciadini et al. 1997b), although there was obviously an immunological reaction to natural bovine BMP in the former, which impaired healing at the later stages of the study (Gao et al. 1997). In rat cranioplasty, natural coral with BMP was superior to natural coral alone (Arnaud et al. 1999).

Tricalcium phosphate has been used as a carrier material either alone or in combination with other materials, especially hydroxyapatite (Urist et al. 1984, Stevenson et al. 1994, Gao et al. 1996a, Boden et al. 1999)

The other materials suggested as carrier materials for BMPs constitute a very heterogeneous group of different materials, including bioactive glass, calcium sulphate, carbon, fibrin sealant and titanium (Lindholm & Gao 1993).