Since the pioneering work of Marshall Urist, who was able
to identify a family of proteins that have the property of osteoinduction,
investigators in many laboratories have shown that bone morphogenetic
proteins (BMPs) will elicit the differentiation of non-committed
stem cells along the line leading to the formation of bone. Not
only are the BMPs able to stimulate progenitor cells to differentiate
and form bone, but in appropriate environments, they can produce
cartilage, tendon, or ligament. Recent data suggest that certain
BMPs may even lead to the partial repair of nerve and kidney. Recombinant forms
of BMPs, particularly BMP2, 4, and 7, have the capability of healing
critical-sized bone defects in rodents, dogs, sheep, and primates
when combined with a carrier of collagen, guanidine-extracted demineralized
bone matrix, hydroxyapatite, or biodegradable polymers. Clinical
trials using a highly concentrated human extract of BMP have shown
promise for the treatment of established non-unions and spine fusion.
Johnson et al. reported in a series of publications that a combination
of internal fixation and implants containing human BMP could lead
to successful union in more than 90% of patients with established
non-unions1. In five patients
with established posterior spinal pseudarthrosis who underwent posterior
spinal fusion with autogenous bone graft augmented with BMP, four went
on to fusion.
Comprehensive clinical trials of recombinant BMIP2 and BMIP7 are
currently underway in Europe and the United States in the areas
of anterior spine fusion, in cages, and in long-bone non-unions
and fractures. Each of these studies has demonstrated that recombinant
BMP was comparable with autogenous bone graft.
With such studies in hand, extending from the lower-level mammals
to humans, why hasn’t the FDA released these growth factors?
Several issues are involved: the under-performance of BMP as one
moves up the mammal chain, the problem of an appropriate carrier,
and the clarification of the pathway to follow through the process
of FDA approval.
Recombinant BMP can lead to a faster rate of union than can autogenous
bone graft in femoral defects of rats, rabbits, and rodents2. However, when the BMPs were applied
to primates, the recombinant BMPs achieved no better success than
did autogenous bone graft3. In
higher mammals, the BMPs may have to be present for a longer period
of time and may have diffused away from the currently used carriers.
Secondly, the number of progenitor cells that are responsive to
BMPs may be more limited in the higher mammals and humans, particularly
under clinical circumstances such as non-unions, than in young rodents.
Thirdly, a family of various forms of BMP is found within bone that
is closely linked to facilitating proteins such as osteocalcin.
When recombinant BMPs are utilized in higher mammals, a single BMP
is chosen and is used in a pharmacological dose rather than in a
physiologic dose. It is still dependent on the recruitment of local
cells, additional BMPs, and other growth factors.
Another problem has been the uncertainty of the ideal type of carrier.
A variety of carriers have been used, including ceramics, collagen,
non-collagenous proteins, biodegradable polymers, and allograft.
Issues of binding, bioavailability, and diffusability and the kind
of carriers have not been resolved. The BMPs that are now being
utilized in human trials are easily diffusable from the collagens,
which hold the BMP for only a short period of time. When carriers
are developed that retain the BMP for a longer period of time, then results
superior to those with autogenous bone graft may be obtained. Concerns
have been raised about the use of single BMP versus a family of
BMPs and the absence of the associated facilitating proteins, the
uncertainty of the pharmacokinetics of the release of the material,
and its duration. Finally, uncertainty remains as to the nature
and number of responding cells at the site of application.
An unforeseen issue that has come to light in recent years regarding
BMP is its role in osteoclast formation. When BMP is provided at
an injury site, there is a stimulation of osteoblast lineage and
also the release of factors that lead to the rapid production of
osteoclasts. The formation of osteoclasts occurs before osteoblast
formation. Large doses of BMP may lead to a wave of resorption that precedes
the appearance of the osteoblasts. Pharmacological agents such as
the bisphosphonates prevent osteoclast formation and activation,
but clinical trials have yet to be performed to demonstrate a control
of the resorptive phase.
Human fibular osteotomy trials using BMP-7 were performed by Geesink
et al.10, who used demineralized
bone matrix as a negative control. It became clear that the demineralized
bone matrix, although it carried less BMP, provided results comparable
with the results of the recombinant BMP. This raised the question
of whether demineralized bone matrix, with its mixture of facilitating
proteins including osteocalcin and osteopontin, may be able to utilize
lower doses or BMP effectively and compete with the recombinant
BMPs. In particular, there is continued research in the area of
modifying allografts into demineralized bone matrix that concentrates
the available BMP but also retains the critical facilitating proteins.
Human clinical trials remain the last barrier. When the BMPs
first came to the attention of the FDA, the agency had no clear
criteria for approval or validation of the efficacy of these growth
factors. Were they a device or should they be tested for biological activity,
and what was the best measure of outcome? Biotechnology companies
providing recombinant BMPs had to work closely with the FDA to establish clinical
trials that demonstrated no toxicity and to develop valid outcome
parameters. Clinical trials that have just been completed were performed
in anterior spine arthrodesis, tibial fractures, and the treatment
of non-unions. Valid outcome parameters had to be established first,
and efficacy then had to be evaluated in clinical trials. This required
a considerable amount of time.
Thus, the road from Marshall Urist’s discovery of the
BMPs, to the development of recombinant BMPs, to the proof of their
efficacy in vitro and in vivo in animal models, and finally to clinical
trials has taken a decade. The BMPs are now in the final approval
stages by the FDA, but the definition of optimal conditions for
their use is not complete. Attention should be directed at developing
appropriate carriers and the mixture, timing, and diffusion characteristics of
the BMPs. Different conditions and combinations may be necessary
for different diseases or anatomical sites. At this time, clinical
trials indicate that the recombinant BMPs are as effective as autogenous
bone grafting. Future research directed at refining the BMP dosage,
time course, release dynamics, and matrix carrier may establish
the superiority of the BMPs over bone graft in humans.