The Effect of Bone Morphogenetic Proteins in Fracture Healing

Abstract

Bone is unique in its selfhealing capacity as it heals without fibrous scar tissue formation. The process resembles the embryonic bone formation in utero, either via the intramembranous or the endochondral route; in fracture healing cells migrate in and differentiate from mesenchymal stem cells into mature osteoblasts or chondroblasts, forming a callus. By remodelling and later modelling bones unite without a scar and in some cases even without radiographic traces. When healing does not occur, the consequences of a fracture may be detrimental to the individual: leaving a true pseudarthrosis, or a malunited bone resulting in poor function and destruction of joints, or infection of the fracture which may lead to permanently draining sinuses or even amputation. Nature’s capacity to heal bones is enormous and therefore our job as surgeons to promote healing is in most cases quite simple. Fractures should be reduced to as close an anatomical position as possible, and immobilised for a certain period of time until enough callus to carry load has been formed. During the 20th century methods for fracture fixation became more sophisticated, and for certain fractures this has changed the outcome dramatically. Our understanding of fracture healing has also improved; the molecular mechanisms and the sequences of healing are partially known. However, we still poorly understand why fractures do not unite in every case, even if a large fracture gap, a poorly reduced fracture or improper immobilisation regularly lead to delayed union or nonunion (1). Since a significant part of fractures does not heal properly, there is still a need for augmentation of the bone healing process. In 1965 Marshall Urist coined the term bone morphogenetic protein when he detected that a protein extract from bone lead to new bone formation in muscle (2). In the 1980’s the amino acid structures of the bone morphogenetic proteins were detected, and the proteins were sequenced and cloned. In the 1990’s, after pre-clinical testing in animals from mouse to non-human primates, recombinant proteins started to be tested in human models. Recombinant human bone morphogenetic protein 2 (rhBMP-2) and recombinant human bone morphogenetic protein 7 (rhBMP-7) or osteogenic protein 1 (OP-1) has so far been used clinically. At present, there are several published clinical studies on the effects of BMPs in bone healing or in delayed unions/ non-unions and several studies have reported the effects in fusion of the lumbar spine (3–5). In this review the clinical studies published on long bone healing will be summarized; one study on fibular osteotomies, one study each on the BMPs used in tibial pseudarthrosis or in allograft pseudarthrosis and two studies on acute tibial fractures (6–10). In the present review BMP denotes the recombinant human proteins and not the term that Urist coined; rhBMPs are not necessarily identical to the proteins produced in the human body as there are posttranslational modifications of the proteins in vivo that do not take place when produced in vitro (phosphorylations and glucosylations of the protein). The amino acid sequence of the rhBMPs is though identical to the ones produced in vivo.