Normal fracture-healing involves a number of osteogenic factors
that are released from bone and the surrounding soft tissues during
the repair process. Osteogenic factors are involved in a number of
processes related to bone formation and bone-remodeling, including
chemotaxis, proliferation, and differentiation of bone-forming and
bone-remodeling cells, blood vessels, nerves, and marrow elements.
Local release of physiologic quantities of these osteogenic factors
is generally sufficient to elicit fracture repair. Considerable effort
has been expended in an attempt to accelerate fracture repair and
to increase the assurance of healing by the exogenous application
of these osteogenic factors. Similar efforts have been made to bridge
critical-sized segmental defects and nonunions with the use of osteogenic
factors. Determining the appropriate delivery vehicle for local application
of these factors has been one of the major limitations to the success
of these therapies. In addition, supraphysiologic doses of osteogenic factors
appear to be required to achieve satisfactory results. Since the
endogenous release of physiologic levels of osteogenic factors does
not require a delivery vehicle for fracture repair, why is a delivery
system required for exogenous delivery and why are the required
doses so high? The answers to these questions may be related to
the type of orthopaedic repair being attempted and the animal model being
studied.
The role of delivery vehicles for osteogenic factors is best
understood in the context of bridging critical-sized defects. These
defects by definition do not heal spontaneously. One of the primary
roles of the delivery vehicle in this type of repair is to maximize the
osteogenic effect of the delivered factors by maintaining these
factors at the site of implantation and optimizing their release
profile. Another important role of the delivery vehicle is to serve
as an osteoconductive matrix for bone-forming cells while maintaining
a space or volume in which bone formation can occur. Maintenance
of an appropriate space to allow bone formation is especially important
in critical-sized defects where there is competition between bone
formation and the surrounding soft tissues encroaching into the
defect. Ideal delivery vehicles should also be biocompatible to
minimize interference with bone induction by excessive inflammatory
reactions. The ideal delivery vehicle should also be biodegradable
to minimize the effects of residual carrier on the biomechanical properties
of the repair. Delivery vehicles for segmental defect repair should
also have the appropriate porosity or granular configuration to
allow for cell infiltration. Commonly used delivery vehicles for
osteogenic factors in segmental repairs include collagen and hyaluronan-based
sponges, pads, pastes, and gels; ceramic blocks and cements; synthetic
polymers; allograft bone; and various combinations of these materials.
The requirement for supraphysiologic amounts of osteogenic factors
can also be best understood in relation to critical-sized defect
repairs, especially in large animal models. One of the reasons these defects
fail to heal may be due to insufficient numbers of responding cells
from the bone envelope and surrounding soft tissues to generate
enough bone to bridge the defect. Although physiologic levels of
osteogenic factors may be present initially, they may not be maintained
for a long enough period to recruit sufficient responding cells throughout
the defect. In addition, upregulation of inhibitors of osteogenic
factors plays a role in localizing bone induction and limits the
long-term effects of these endogenous factors. As a result, soft tissues
interpose between the bone ends and prevent bridging of the defect.
The addition of supraphysiologic levels of osteogenic factors to
delivery vehicles may be required to maintain physiologic levels
of these factors for a sufficient period of time to stimulate enough
responding cells throughout the defect to support bridging. As new
delivery vehicles with more optimal release profiles are developed,
the need for excessive starting doses of osteogenic factors may
be decreased.
The requirement for supraphysiologic levels of osteogenic factors
to bridge critical-sized defects in large animals introduces the
role of animal models to this discussion. Most studies have indicated
that higher doses of osteogenic factors are required for large animal
models compared with those for rodent and rabbit models. Even higher
doses appear to be required in nonhuman primates. Interestingly, critical-sized
defects in rodents can be bridged with osteoconductive matrices
that do not contain any exogenous osteogenic factors. Sufficient
endogenous factors appear to be present to allow bridging if the
osteoconductive matrix maintains a space or volume for bone formation
and prevents soft-tissue collapse into the defect. Critical-sized
defects in rodents can also be bridged with the addition of osteogenic
factors delivered to the site with viral vectors, transfected cells
containing genes expressing osteogenic factors, and direct implantation
of cDNA encoding for the osteogenic factors. There are also a few
reports of bridging critical-sized defects in rabbits with use of
some of these modalities. When osteogenic factors are evaluated
in these small animal models, much lower doses are required to bridge
the defects compared with those in large animal models. These findings
have led to a general consensus that rodent and rabbit models may
be too permissive with respect to evaluation of osteogenic factors
for bone induction in people. The rationale for this enhanced ability
to bridge critical-sized defects in rodents and rabbits compared
with larger animal models may be related to an increase in the number
of responding cells in the bone and soft-tissue elements and a more
rapid rate of bone formation. Similar arguments have been used to
explain the dose escalation required in large animal models and
nonhuman primates. Higher initial doses are thought to be required
in large animal models to maintain physiologic levels of osteogenic factors
for the longer length of time required to recruit sufficient cells
into the defect compared with those in small animal models. As a
result of these differences, extrapolation of results from rodents
and rabbits to larger animals, including nonhuman primates and humans,
may not be warranted. Paradoxically, failure of some delivery vehicles
in rodents and rabbits may be the result of prolonged residence
time of the vehicle interfering with rapid bone formation in these
animal models. The prolonged residence times of these delivery vehicles
may be efficacious in larger animal models where bone formation
may not be as rapid. The location of the critical-sized defects
used in animal models is also important. In some instances, defects in
intramembranous bone can be bridged with osteogenic factor-vehicle
combinations that do not work as well in defects in endochondrally
derived bones.
In contrast to critical-sized defects, fractures generally heal
spontaneously in response to endogenous osteogenic factors. However,
this process does not appear to be optimized on the basis of the
numerous studies that have demonstrated acceleration of fracture-healing
in response to osteoconductive matrices and exogenous osteogenic
factors compared with untreated controls. These results suggest
that more rapid and larger callus formation can be achieved compared
with the normal healing process. In addition to an increase in the
periosteal and endosteal response, an important aspect of accelerated
fracture-healing appears to be an increase in the contribution of
the soft tissues surrounding the bone to the repair process. This
increased soft-tissue response appears to be achieved by an increase in
the number of responding cells from the soft-tissue envelope, including
muscle, fascia, vascular pericytes, and nerves.
The role of delivery vehicles for osteogenic factors is less
well understood in relation to accelerating fracture repair compared
with bridging critical-sized defects. The ideal characteristics
of delivery vehicles for osteogenic factors for acceleration of fracture
repair share some of the features described above for segmental
defect repair. Ideal delivery vehicles should be biocompatible and
biodegradable. As was the case in segmental defect repair, ideal
vehicles should also optimize release of the osteogenic factor at
the fracture site. However, unlike segmental defect repair, there
is less need for the vehicle to maintain a space or volume for bone formation.
In fact, vehicles for accelerating fracture repair must be rapidly
degraded so that there is minimal interference with normal fracture
repair. The combination of optimal factor release combined with
rapid degradation is a difficult goal to achieve. In addition, these
delivery vehicles may also need to be injectable through a 16 to
18-gauge needle in order to allow percutaneous treatment of closed
fractures. Commonly used delivery vehicles for osteogenic factors
to accelerate fracture-healing include naturally derived polymers
such as collagen, hyaluronan, chitosan, and fibrin; synthetic polymers;
ceramic materials including injectable calcium phosphate cements;
and various combinations of these materials.
As was the case in segmental repairs, osteoconductive matrices
that do not contain osteogenic factors have been used to accelerate
healing of diaphyseal fractures in rodents. With the exception of
bone cements used to enhance unions in metaphyseal fractures, there
are few reports of acceleration of fracture-healing in large animal
models with osteoconductive matrices. The role of the bone cements in
metaphyseal fractures is more closely related to segmental defect
repair. The vehicles are used to fill gaps in the repair or are
used to support the articular surface. These materials are then
slowly resorbed over time and replaced by bone. Successful acceleration
of healing has also been achieved with use of osteogenic factors
delivered in viral vectors, transfected cells containing genes expressing
osteogenic factors, in rodents and rabbits. Osteogenic factors injected
in formulation buffer have been used to accelerate osteotomy-healing
and fracture-healing in a number of animal models. These models include
rats, rabbits, dogs, and sheep. The use of formulation buffer as
a delivery vehicle would be ideal since there would be minimal interference with
cell infiltration at the fracture site. However, there are no reports,
as far as I know, of successful acceleration of diaphyseal osteotomy-healing
or fracture-healing with use of osteogenic factors delivered by
injection in formulation buffer in nonhuman primates or people.
Successful acceleration of osteotomy-healing has been reported in
a wide variety of small and large animal models with use of osteogenic
factors delivered in hyaluronan gels, collagen pastes, and calcium
phosphate cements. Hyaluronan and calcium phosphate cements have also
been used successfully as delivery vehicles in nonhuman primate
osteotomy models. Several clinical trials are currently being conducted
to evaluate osteoconductive matrices and osteogenic factor-vehicle
combinations in metaphyseal fractures and long-bone fractures in
people.
Dose escalation for osteogenic factors has also been observed
in large-animal and nonhuman-primate osteotomy and fracture-healing
models compared with small animal models. As was the case in segmental
repairs, the explanation for the discrepancy between nonhuman primates
and people compared with other animal models has been attributed
to differences in numbers of responding cells, rates of fracture-healing,
and residence time of osteogenic factors. Higher initial doses are
thought to be required in large animal models of fracture repair
to maintain physiologic levels of osteogenic factors for the longer
period of time required to recruit sufficient cells into the repair
compared with those required in small animal models. The development of
new delivery vehicles with optimal release profiles may decrease
the need for high initial doses of osteogenic factors.
The use of osteotomy compared with closed-fracture animal models
also has generated some controversy in the study of delivery vehicles
for osteogenic factors used to accelerate fracture repair. The major
advantage of the use of osteotomies is the ability to standardize
the bone and soft-tissue injury between animals used in a study.
The major disadvantage of the osteotomy model compared with the
closed-fracture model is the rate of healing. In general, most osteotomy
models represent delayed healing compared with nondisplaced closed-fracture
models. There is also some controversy as to the extent of the associated
soft-tissue damage incurred in the osteotomy model compared with
that in a closed-fracture model. The major advantage of the closed-fracture
model is the better approximation to closed fractures in people
with respect to fracture configuration and associated soft-tissue
injury. However, depending on the method used, the fracture configurations
between animals can be variable, as can the degree of soft-tissue
injury. In addition, most closed fractures in animal models heal
at a much more rapid rate than do closed fractures in people. This
is especially true when the rate of healing of closed tibial fractures
in animals is compared with that in people. Accessing the efficacy
of osteogenic factor-delivery vehicles in these rapidly healing
models may not be relevant for treating the same fractures in people.
The use of veterinary fracture clinical cases, which can be identical
to human fracture clinical cases, may also be limited by this difference
in the rate of fracture-healing. Animal models of diaphyseal fracture-healing
also may not apply to metaphyseal fracture-healing. Differences
in metaphyseal fracture-healing may be related to the increased
number of responding cells residing in trabecular bone compared
with cortical bone and to the soft-tissue elements associated with
diaphyseal fractures.
Given the above discussion, delivery systems for osteogenic factors
will most likely be required to achieve a significant improvement
in segmental defect repair and to accelerate fracture-healing in humans.
Osteoconductive matrices without osteogenic factors may also have
a role in improving the repair of metaphyseal fractures. Animal
models are critical to establish safety and toxicology data prior to
initiating clinical trials in humans. These models can also be very
valuable in comparing the release kinetics of osteogenic factors
from different delivery vehicles in vivo. The use of animal models
to establish efficacy and dose-ranging for human clinical trials
needs to be evaluated carefully. There is general agreement that
osteogenic factor-delivery vehicles that do not work in rodents
and rabbits will most likely not work in larger animal models. If these
combinations do not work in larger animal models, they will most
likely not work in humans. Conversely, successful testing in small
animal models may not be predictive of performance in larger animal
models or humans. However, the degree of comfort that a successful
outcome will be predictive in people usually increases with success
in large animal models. Of the large animal models, nonhuman primates
probably represent the closest approximation to humans. As is the
case with all animal models, studies involving nonhuman primates
should be designed with great care such that a minimum number of
animals are used while retaining sufficient statistical power to
draw the appropriate conclusions from the results. Most importantly,
there is no guarantee that combinations that work in small and large
animal models, including nonhuman primates, will work in people. That
is what human clinical trials are for.