- Email: [email protected]
Healing of segmental bone defects with granular porous hydroxyapatite augmented with recombinant human osteogenic protein-1 or autologous bone marrow
ELSEVIER
Journal of Orthopaedic Research
Journal 01' Ortliopoedic Research 21 ( 2 0 0 3 ) S21 52X \\!AM,
elsevier coinlloc'itelorthres
Healing of segmental bone defects with granular porous hydroxyapatite augmented with recombinant human osteogenic protein- 1 or autologous bone marrow Frank C. den Boer 'I,*, Burkhard W. Wippermann b, Taco J. Blokhuis ", Peter Patka ", Fred C. Bakker 'I, Henk J.Th.M. Haarman
Abstract
Hydroxyapatite is a synthetic bone graft, which is used for the treatment of bone defects and nonunions. However, it is a rather inert material with no or little intrinsic osteoinductive activity. Recombinant human osteogenic protein- I (rhOP-I) is a very potent biological agent. that enhances osteogenesis during bone repair. Bone niiirrow contains inesenchynial stcm cells. which are capable of new bone formation. Biosynthetic bone grafts wcrc created by the addition of rhOP-I or bone marrow to granular porous hydroxyapatite. The performancc of these grafts was tested i n ;i sheep model and compared to the results of autograft, which is clinically the standard treatment of bonc dcfccts and nonunions. A 3 cni segmental bone dcfcct was made i n the tibia and fixed with an intcrlocking intraiiiedullary nail. There were five treatment groups: i i o implant ( n =: 6 ) . autograft ( 1 7 = 8). liydroxyapatite alone ( n = 8). liydroxyapatite loaded with rhOP-I ( i i = 8). and hydroxyapatite loaded with autologous bone marrow ( n = 8). At 12 weeks, healing of the dcfcct was evaluated with radiographs, a torsional test to F~ilure,and histological examination of longitudinal sections through the defect. Torsional strength and stillness of the healing tibiae were about two to three times higher for autograft and hydroxyapatite plus rhOP-I or bone niiirrow compared to hydroxyapatite alone and empty defects. The mean values of both combination groups werc comparable t o those of autograft. Thew wcre more unions in defects with hydroxyapatite plus rhOP-l than in defects with hydroxyapatite alone. Although the diflcrences were not significant, histological examination revealed that there was inore oftcn bony bridging of the defect in both combination groups and the autograft group than in the group with hydroxyapatite alone. Healing of bone defects. treated with poroiis hydroxyapatite. can be enhanced by thc addition of rhOP-l or autologous bone marrow. The results of these composite biosynthctic grafts are equivalent to those of autograft. 0 2002 O r t h o p x d i c Rescarcli Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction
In trauma surgery, bone loss creates a major technical and biological problem. Fractures in cancellous bone, such as i n the proximal humerus, thc distal radius, or the tibia1 plateau, often lead to impaction of bone and consequently a defect results after rcduction. Otheicauses of traumatic bone loss are fresh highly coniniinu ted fi-rict it res mid ;it 1-0ph ic ti oti u n i oti s in the d i a ph ys i s of long bones. Autologous bone grafting is widely accepted as the gold standard for the treatment of bone .. .. .. .
..
CorrcsponcIing a u t h o r . T ~ I . +3 : I -20-444-0268: fits: +3 I -20-4440274. l~-iii(ri/ i i t / d w r t : fl.aiikden boer.~~~liotrn;iil.coiii (t'.C. dcii Boer.).
dcfccts and nonunions [20]. However, autologous bone grafting has some serious drawbacks, such as a prolonged opcrntion time [6] and donor site morbidity in about 10-30'%, of the cases [7,29]. I n order to avoid the problems associated with autologous bone grafting, there has been a continuous interest in the use of synthetic bone grafts during the past decades. The most frequently used synthetic bone gr a fts are calcium phosphate ceramics, such as hydroxyapatite, which is known for its excellent biocompatibility [3.16,17]. The intcrconnectivc porous structure of many hydroxyapatite based ceramics acts a s a scaffold and allows the ingrowth of newly formed bone into the pores, which is known a s ostcoconduction [5,19,2123.26,30]. In prospective clinical trials. liydroxyapatite
ncc I . i t l All rights i-e\ei-\ecl. 07.X-O2h6/03/$ - see I'roiit matter. (c) 2002 0rtliop;wdic Rtrearch Societ). Publihed bq E l w icr dol: 10. ~ ~ ~ ~ 6 / ~ ~ ~ ~ ~ 6 - ~ ~ ~ ( , ~ ( ~ ~ ~ ) ~ ~ ~ ~ ~ ~ ~ ~ - ~
522
F.C. diw Boer r t al. I Jourrrul of Orihupuedic Restarch 21 (20031 521 528
ceramics have been used successfully in bone defects, caused by fractures [6,7]. However, hydroxyapatite is a rather inert material with no or little intrinsic osteoinductive activity 1171. In order to enhance new bone formation, growth factors or osteogenic cells are added to hydroxyapatite [4,5,20]. Bone morphogenetic proteins (BMPs) are a group of growth and differentiation factors, that play an important role in embryonic morphogenesis of skeletal and nonskeletal tissues and in osteogenic differentiation during bone repair in postnatal life [31]. One of the BMPs is osteogenic protein-1 (OP-1 or BMP-7), which is manufactured with recombinant DNA techniques [lo]. In vitro studies have shown that recombinant human OP-1 (rhOP- 1) stimulates proliferation and osteogenic differentiation of mesenchymal cells [ 1,241. In criticalsized segmental bone defects in rabbits, dogs, monkeys, and humans, rhOP- 1 induced healing within 2-3 months [8,9,11,14]. These data indicate that rhOP-1 has great potential as a therapeutic agent in bone repair. Bone marrow contains mesenchymal stem cells that possess the capacity to differentiate into osteoblasts [4,13]. Bone marrow is generally considered as an important source of cells, responsible for new bone formation during bone repair. It has been demonstrated in animal models, that fresh autologous bone marrow induces healing of segmental bone defects [15,27]. In this study, composite biosynthetic bone grafts were obtained by the addition of rhOP-1 or autologous bone marrow to granular porous hydroxyapatite. These composite grafts were tested in an established segmental bone defect model in sheep [12]. In order to assess the efficacy of the added rhOP-1 or bone marrow, the results were compared to those of pure granular porous hydroxyapatite. In addition, the performance of the two composite biosynthetic grafts was compared to that of autograft, the gold standard in bone grafting.
Materials and methods Surgicul iwocidwes und .I rucl,: tic,.sign
This study was approved by the institutional review boards for animal experiments. Thirty-nine adult female sheep (thirty-one NorthHolland sheep and eight black-Faced sheep) were used. The mean weight ( z t standard deviation) was 54.2 & 7.6 kg. The animal model, consisting of a 3 cm segmental tibial defect fixed with an interlocking intramedullary nail. has been described previously [12]. In short, the surgical technique was as follows. The left tibia of each animal was operated upon. A 3 cm segmental defect, centred between the innermost interlocking holes of the intramedullary nail, was created by performing two osteotomies with an oscillating saw under constant cooling with saline. The periosteum was not resected. A custom-made A 0 tibial nail (Synthes, Bochum, Germany) was used for fixation. The parts of the tibia, proximal and distal to the defect, were separately reamed. Then. the nail was inserted through the tibial head with the knee in a flexed position. Proximal interlocking was performed, using the A 0 proximal aiming device. A custom-made distal aiming device was used for distal interlocking. Proximally and distally, all three holes were used for interlocking. To reduce differences in outcome between
anirnals, because of differences in the number of osteogenic cells present in the bone marrow at the defect site, all bone marrow in the defect and the medullary cavity was removed by copious irrigation and suction both before and after insertion of the nail. According to the method of treatment of the bone defect, there were five experimental groups: ( I ) empty defects (7 animals), (2) 10 ml morseled autologous corticocancellous bone, harvested from the ipsilateral posterior iliac crest (8 animals), (3) 10 ml granular porous hydroxyapatite alone (8 animals), (4) 10 ml granular porous hydroxyapatite mixed with 2.5 mg rhOP-l (8 animals), and (5) 10 ml granular porous hydroxyapatite mixed with 10 ml autologous bone marrow (8 animals). The soft tissues were closed meticulously over the defect with Vicryl, which ensured containment of the graft materials. As an analgesic 0.5 mg buprenorfine was administered intramuscularly twice daily during the first two days. One sheep with an empty defect sustained a deep infection and this animal was excluded from further analysis. which left six animals in this group. The used calcium Dhosuhate ceramic (Endobon'*, Merck Biomaterials, Darmstadt, Germany) is produced from bovine cancellous bone by a sintering process. This is a manufacturing process under high pressure and at a high temperature (1250 "C), by which the crystallinity of the native hydroxyapatite increases and fusion of crystals occurs. thereby increasing the solidity of the material [2]. Endobon" is an inorganic material. consisting mainly of crystalline hydroxyapatite in the form of granules. sized 2.8 5.6 mm. with an interconnecting pore structure identical to the original cancellous bone. The porosity is 30-80'%1(mean 55%) and the pore size ranges from 100 to 1500 pm with a mean of 450 pm. Prior to surgery. the hydroxyapatite granules were sterilized by dryheating at 180 "C for 30 min. The rhOP-l was bound to bovine bone type I collagen carrier. which is a particulate substance with a size of 75-425 pm. In each animal 2.5 mg rhOP-I and 1.0 g collagen carrier (OP-1 Device, Stryker Biotech, Natick, Massachusetts) was reconstituted with sterile saline, mixed with the hydroxyapatite granules, and implanted in the defect. The 2.5 mg rhOP-1 dose was based on studies with segmental bone defects in monkeys and humans. in which the defects were treated with 1-2.5 mg r-hOP-1 [I I , 141. Autologous bone marrow was aspirated from the ipsilateral posterior iliac crest with a Jamshidi needle on a syringe. A total volume of 10 ml was obtained by three punctures and aspirations of about 3 4 ml in different directions at one site. The bone marrow was mixed with the hydroxyapatite granules and allowed to clot. As part of this cooperative research project. the operations and the evaluation of healing of all animals in the hydroxyapatite plus bone marrow group were performed at the Medical School in Hannover, Germany. Because North-Holland sheep were not available in Germany, black-faced sheep were used in this group: in the other groups, North-Holland sheep were used. The mean weight for the animals with hydroxyapatite plus bone marrow was 59.0 kg and comparable to the mean weight in the other groups, which ranged from 50.0 to 55.4 kg. One-way analysis of variance showed no statistical difference in body weight between the treatment groups ( p = 0.14). The operation technique, the operator (principal author), and the analysis techniques of healing were the same in the hydroxyapatite plus bone marrow group and the other groups. Although the radiographic and biomechanical data of the control groups with empty defects and autologous bone have been described previously [12], these data are included for comparison with the hydroxyapatite groups, which is relevant to the research questions. After 12 weeks all sheep were killed, both tibiae were explanted, and dissected free of soft tissues. The nail was removed and immediately thereafter, radiographical and biomechanical analysis was performed. This 12-weeks end-point was chosen, because in large animal segmental long bone defect models, 12 weeks is a usual and suitable interval for the detection of differences in bone repair among different treatments [8,18,30]. 1
.
Ratliogruphic eouluution
Anteroposterior and lateral radiographs of the left tibia were taken immediately after surgery to evaluate the position of the nail, the locking screws, and the graft material. After 12 weeks, contact radiographs were taken of the explanted tibiae, which were stripped of all soft tissues. Healing of the defect was evaluated by two clinicians, who were unaware of the treatment group. The extent of healing was gra-
F.C. den Boer et ul. I Joumml
of Orthopuedic Reseurcli 21 (2003) 521
ded on a scale of 0 4 , indicating the number of sides, relatively to the medullary canal on the radiographs in two directions, with a uniting bone callus. T o prevent the need for frequent sedation, sequential radiographs in the intervening period till 12 weeks were not made.
528
523
in the number of sides with bony bridging, as assessed by histology. were evaluated with the Kruskal-Wallis test. For all statistical tests a p value of less than 0.05 was considered to be significant. All statistical analyses were performed with the software program SPSS for Windows, version 7.5 (SPSS Inc., Chicago, IL. USA).
Bioiiic~cl7uniccrltesting
Immediately after explantation, a torsional test to failure was performed on both tibiae of each animal. The ends of the experimental tibia were embedded in dental plaster (Vel-mix Stone, Kerr GmbH, Karlsruhe. Germany) in such a way, that the nonembedded portion comprised the 3 cm defect and 1 cm of the diaphysk proximal and distal to it. Because all holes of the locking bolts were embedded, failure through these holes during the test was prevented and the mechanical properties of the defect only determined the outcome of the test. The nonembedded portion in the contralateral intact tibia consisted of an identical 5 cm segment of the shaft, located at the same level. With a custom-built mechanical testing machine the bones were tested to failure by exorotating the proximal end at 20" per minute. Torque was continuously measured at the fixed distal bone end. Torque angular displacement curves were generated, from which the torsional strength (torque at failure) and stiffness were calculated. T o reduce errors associated with geometric or size differences among animals in different groups, torsional strength and stiffness were expressed as a percentage of the value of the contralateral intact tibia. A nonunion was defined as a bone defect with a torsional strength or stiffness, or both. of less than 20% that of the contralateral intact bone [12]. Below the 20'%1 level, the torque^ angular displacement curves showed a slow nonlinear increase in torque indicating plastic deformation. which is consistent with the presence of a fibrous nonunion. N o other criteria, such as the radiographic appearance, were used for essment of nonunions. As shown previously, radiographic evaluation is unreliable for the assessment of nonunions, because new bone formation in the defect can sometimes give a false impression of a uniting bone callus, probably as a consequence of overprojection [12]. The torsional test results were not available for three animals; one in the autogi-aft group due to failure of the recording during the test and two in the hydroxyapatite plus bone marrow group due to an error in the processing of the bones after explantation. IIi.stologiccil tmuljsis
The specimens, containing the 3 cm segmental defect and I cm proximal and distal cortical bone. adjacent to the defect, were fixed in alcohol 70'%,.Longitudinal sections with a thickness of about 2 mm in the central portion of the bone in the frontal plane were sawn with use of a diamond-impregnated band saw (Exakt Apparatebau, Norderstedt, Germany). These undecalcified sections were dehydrated in ascending grades of alcohol and embedded in Technovit (Technovit 7200 VLC, Kulzer GnibH. Wehrheim, Germany). Then 50 pm sections were made, using a sawing and grinding technique (Exakt Apparatebau, Hamburg, Germany). Of each animal one section was stained with Goldner's trichrome stain and one with toluidine blue 0.2'%1.The sections were examined with use of a light microscope. The amount of fibrous tissue and new bone formation and the presence of remodeling were qualitatively assessed. The presence of new lamellar bone and of osteoclasts along the trabeculae of the newly formed woven bone were considered a s signs of remodeling. Bony bridging was defined as a continuous connection with newly formed bone between the proximal and distal end of the defect. Bony bridging was scored as none, one side (medial or lateral), or two sides (medial and lateral), relatively to the medullary canal. Sturitr ic.u
One-way analysis of variance was used to test for differences in the biornechanical data among different treatments, followed by pairwise comparisons between the groups with the post-hoc Duncan test. The numbers of nonunions in the experimental groups were compared with Fisher's exact test and to minimize the influence of multiple comparisons on the results, only a limited number of comparisons were made: hydroxyapatite alone versus empty defects and both combination groups versus hydroxyapatite alone or autograft. Differences among the treatment groups in the median radiographic score of healing and
Results Radiography On the twelve-week radiographs, there was more new bone formation in both combination groups and the autograft group compared to hydroxyapatite alone (Fig. 1). The median radiographic score of healing at this time was 1.5 for empty defects, 2.5 for autograft, 1.0 for hydroxyapatite alone, 2.0 for hydroxyapatite plus rhOP1, and 3.0 for hydroxyapatite plus bone marrow. With the numbers of animals available, there were no significant differences in the radiographic score among the groups 0, = 0.28). Biornechanical tests Torsional strength was comparable for empty defects and hydroxyapatite alone (Fig. 2A). For hydroxyapatite plus rhOP-1, hydroxyapatite plus bone marrow, and autograft, torsional strength was about two to three times higher compared to hydroxyapatite alone or empty defects. There was no difference between the autograft group and both combination groups. For torsional stiffness, the results were comparable (Fig. 2B). According to the biomechanical criterium, there were more nonunions in the empty defect group and the hydroxyapatite alone group compared to autograft, hydroxyapatite plus rhOP-1, and hydroxyapatite plus bone marrow (Table 1). The difference between hydroxyapatite alone and hydroxyapatite plus rhOP-1 was significant ( p = 0.03). Histology In general, the histological findings were in accordance with the results of the radiographs and the biomechanical tests. In all animals, the defect was filled with newly formed woven bone and fibrous tissue. The newly formed bone in the groups with hydroxyapatite was in direct contact with the ceramic, both around the granules and inside the interconnecting pores of the granules, indicating pore ingrowth. By this process of perigranular and intragranular bone ingrowth, the hydroxyapatite granules became incorporated in the newly formed bone. Remodeling of the newly formed woven bone into lamellar bone was present in all cases. Qualitative judgment of the histological sections revealed that in defects treated with hydroxyapatite plus rhOP- 1, hydroxyapatite plus bone marrow, or autograft, the
524
Fig. 1. Anteroposterior and 1;itersl contact radiographs altcr reiiioviil 01' the soft tissues and the nail tit I2 weeks. ( A ) Hydroxyapatitc alone. There ia no new bone formation. the hydroxyapatite granules are still clcai-ly cisible. and the radiographic scoi-c is 7ero. The biomechanical test results are conaistent with ii nonunion. ( B ) Hydroxyapatite plus rhOP- I . As a result ol' abundant new bone formation, the hydroxyapatite granules have become blurred. There is II uniting bone callus at all sidcs around the niedullary canal and the radiographic scow is lour. ( C ) Hydroxyapatite plus autologous bone marrow. Again. there is abundonl new bone Ibrmation with a uniting hone callus a t all four sides and B radiographic score of four.
6o
forsional strength (%)
(A)
.
torsional stiffness (%)
**
***
50 40 -
30
-
70r 60 50
(6)
i
T
~
40 -
20 10
-
10
-
0
Fig. 2. Torsional strength ( A ) and torsional stiflness ( B ) a t 12 weeks. The reaulls tire shown a s a percentage o l ihc value 01' the intact contralateral tibia. The bars indicate the mean values, and the I-bars indicate the standard errors of the mean. ABG: autologous bone grafting, HA: hydroxyapatite. rhOP-I: recombinant human osteogenic protein-I. and BM: bone marrow. (A) Analysis of variance: 11 = 0.02. Post-hoc Duncan test: (*) p < 0.05 vs. HA: ( I * ) p <: 0.05 vs. HA and p < 0.05 vs. empty; (***I 11 < 0.05 vs. HA. ( B ) Analysis o f variance: p = 0.02. Post-hoc Duncan test: (*) p < 0.05 vs. HA iind / J < 0.05 vs. empty.
amount of newly formcd bone was larger and fibrous tissue formation was less than in defects treated with hydroxyapatite alone, (Fig. 3 ) . In some animals with hydroxyapatite alone there was only minimal new bone formation and almost the whole defect consisted of remaining hydroxyapatite granules embedded in fibrous tissue. Correspondingly, bony bridging was observed more often in thc groups with hydroxyapatite plus rhOP-1, hydroxyapatite plus bone marrow, or autograft than in the group with hydroxyapatite alone (Table 2). Howevcr, these difrerences were not significant 0,= 0.25).
Discussion
In traumatic bone dcfects, a bone graft substitute should contain a material with some structural integrity and a space-occupying effect in order to maintain a proper reduction and continuity between the bone fragmcnts during the operation and to enable an adequate osteosynthesis to be performed. It could be argued that thc brittleness of granular porous hydroxyapatite, as used in this study, would make this material unsuitable for this purpose. However, in a traumatic bone defect, a space-occupying effect and little structural
F.C clcw Boer et ul. I Journul of Orthopueiiic Reseurch 21 (2003) 521-528 Table 1 The numbers of unions and nonunions at 12 weeks according to the biomechanical test criteriumd Treatment group
Union
Nonunion
Empty defects Autograft' Hydroxyapatite Hydroxyapatite Hydroxyapatite
2 6 3 8 5
4
+ rhOP-1' + bone marrow'
1 5 0 1
" A nonunion was defined as a bone defect with a torsional strength or stiffness, or both, of less than 20% that of the contralateral intact bone. bThe biomechanical test data were available and thus the presence of a union or nonunion could be assessed for seven of the eight animals in the autograft group and six of the eight animals in the hydroxyapatite plus bone marrow group (see Materials and Methods). rhOP-1 = recombinant human osteogenic protein-I, p = 0.03 compared to hydroxyapatite alone (Fisher's exact test).
support is sufficient, because the major part of the stability is provided by the osteosynthesis and the involved bone will be kept unloaded until healing has been achieved. Although it has been shown that rhOP-1 alone or autologous bone marrow alone are efficient adjuvants in the healing of segmental bone defects [8,9,11,14, 15,271, these materials do not provide structural support. rhOP-1 and bone marrow are (semi-)liquid materials without structural integrity or a space-occupying
525
Table 2 The numbers of animals according to the number of sides wi[h bony bridging, assessed by histological examination of a frontal section, centrally in the defect, at 12 weeks Treatment group
~~
No bridging
One side Two of bridging sides of bridging
2 2 6 2 3
4
~
Empty defects Autograft Hydroxyapatite Hydroxyapatite Hydroxyapatite "rhOP-l
+ rhOP-I" + bone marrow
= recombinant
6 2
6 3
-7
human osteogenic protein-I.
effect. Therefore, no groups with rhOP-1 or bone marrow alone were tested and granular porous hydroxyapatite was taken as the basic material for several bone graft substitutes. In several experimental studies, the combination of hydroxyapatite and BMPs has been tested [ 19,23,25,26]. However, ectopic sites, cancellous bone and calvarial defects, or small animal models were used. In this study, the combination of hydroxyapatite and a BMP was evaluated in a large animal segmental cortical bone defect model. Good results in this model would indicate a highly effective bone graft substitute, which could be useful in many situations of traumatic bone loss
Fig. 3 . Photomicrographs of longitudinal sections of the bone defect in the frontal plane (Goldner's trichrome stain; original magnification. x I ) . The dense tissue at the top and the bottom is the original cortex at the proximal and distal boundary of the defect. The gap between the left and right side is the medullary cavity after removal of the nail. (A) Animal treated with granular porous hydroxyapatite plus rhOP-I, showing abundant new bone formation at both sides. At the lateral (right) side, bone has completely bridged the defect, connecting its cut margins. The score for bony bridging is I . The crack a t the lateral side is caused by the preceding mechanical test. The dark porous particles are the hydroxyapatite granules incorporated in the newly formed bone (arrows). (B) Animal treated with hydroxyapatite alone. There is very little new bone formation at the cut bone ends at the lateral (right) side and in the centre of the defect at the medial (left) side. The rest of the defect is filled with fibrous tissue, surrounding the hydroxyapatite granules (arrows). The score for bony bridging is 0. This animal had a nonunion, which was confirmed in the biomechanical test.
encountered in clinical practise. In addition, the combinations of hydroxyapatite with BMP or bone marrow, both of which were included in the present study, have never been compared to each other. The long bone segmental defect model in the current study is not a consistent nonunion model, since some of the empty defects healed. This can probably be attributed to the retained periosteum, which acts as a source of supply of osteoprogenitor cells [12]. Even in this model, which is not strenuous, different healing responses were found with different graft materials. A disadvantage is that healing was evaluated at only one time-point, i.e. 12 weeks. As a consequence, no conclusions can be drawn about the healing process in the course of the time. For a complete biomechanical and histological evaluation at different times, a lot more animals would be needed for a sufficient number of animals in each group at each time, which is undesirable from an ethical point of view. However, different healing responses with several graft materials, even at one time, indicate which graft materials are most efficient and suitable as bone graft substitutes. In the present study, implantation of pure hydroxyapatite yielded limited new bone formation and the nonunion rate was high and comparable to that of empty defects. In several other animal models with segmental bone defects, the performance of pure hydroxyapatite was comparatively bad [.5,18,21,22]. This is in agreement with the general assumption that hydroxyapatite itself has no osteoinductive activity [17]. Moreover, this ceramic does not contain osteogenic cells. Thus, the bone healing process in defects with pure hydroxyapatite was largely dependent on the osteoconductive properties of the material. However, the granular form of the hydroxyapatite prevented a tight fixation of the ceramic to the adjacent bone, which seems to be a prerequisite for optimal osteoconduction and effective healing [17]. So, the ceramic did not add anything to the regenerative process in the present model. In contrast, efficient healing with much new bone formation and a union rate of loo'% was found, when granular porous hydroxyapatite was combined with rhOP- 1, Calcium phosphate ceramics are suitable carrier materials for BMPs [ 161 and several animal experiments dealing with this topic have been reported. In rodents, purified BMP combined with porous hydroxyapatite induced bone formation in ectopic sites, while negative controls with hydroxyapatite alone did not [19,26]. In cancellous bone defects in rabbits and in calvarial defects in baboons, the addition of purified BMP to hydroxyapatite resulted in significantly more new bone formation [ 193.251. A major disadvantage of purified BMP is the presence of contaminating proteins, that could incite immunogenic responses and carry the risk of disease transmission [X,9, I I , 191. These problems are
prevented by use of rhBMPs, such as rhOP-I, which was used in this study. Similar to almost all previous experimental work with rhOP1, bovine bone type I collagen was used as a carrier material. There are no exact data available on the pharmacokinetics of rhOP- 1 released from different carrier systems. For rhBMP-2, it has been shown that this protein exhibits a sustained release from bovine type 1 collagen and that there is a strong binding affinity to hydroxyapatite mineral with little release after ectopic implantation in rats [28]. Because different BMPs share common structural features [10,31], it could be that rhOP-l in the current model was gradually released from the collagen particles and subsequently retained at the defect site by the hydroxyapatite particles, enabling the protein to exert its osteoinductive effect. Regardless of the exact local pharmacokinetics, the release pattern of rhOP-1 from the collagen particles and the hydroxyapatite granules was consistent with the requirements of the healing bone, as became evident by efficient healing of the defects in this group. It could be argued that the efficacy of rhOP-1 itself has not been proven, because the control group with hydroxyapatite alone did not contain the collagen carrier. However, in experimental segmental bone defects in rabbits, monkeys and humans, the results of the controls with bovine bone type I collagen carrier alone were similar to those of empty control defects with a fibrous nonunion in all cases [9,11.14]. Therefore, it is unlikely that the collagen carrier had an important influence on the bone healing process in the present model. Comparable to the effect of added rhOP-I, the addition of fresh autologous bone marrow to hydroxyapatite resulted in an enhancement of bone healing, which was equivalent to healing with autograft. In the first place, the favourable effect of bone marrow can be attributed to the supply of osteogenic mesenchymal stem cells [4,13,20]. Secondly, the clot associated with bone marrow contains growth factors and the bone marrow cells themselves secrete inductive factors, which stimulate osteogenesis [.5,22]. In other studies with segmental bone defects, an improvement of healing with the combination of hydroxyapatite and bone marrow compared to hydroxyapatite alone was found as well [18,22]. In a 2 cm segmental defect in the sheep tibia, the biomechanical test results were similar for autograft and porous hydroxyapatite plus bone marrow at three and six months [30].In a large prospective randomized clinical trial with fresh traumatic bone defects, the efficacy of the combination of a hydroxyapatite based ceramic and autologous bone marrow has also been demonstrated with results comparable to those of autograft [7]. The results of the present study indicate that rhOP-1 and autologous bone marrow are equally effective as an augmentation of porous hydroxyapatite. rhOP-1 is always off-the-shelf available and can easily be mixed with
I? C. h i Borr
(,I
(11. I Jourriul
of Ortliopurrlic Resrur-cli 21 (2003) 52I-528
granular porous hydroxyapatite; however, the application of recombinant growth factors is rather expensive. Bone marrow needs to be harvested intraoperatively and it is not always possible to obtain a sufficient amount of osteogenic cells. The number of osteoprogenitor cells in bone marrow is relatively small [4,20] and declines with aging [ 131. Conceivably, it is sometimes impossible to harvest a sample of bone marrow with sufficient osteogenic activity in old osteoporotic patients. However, the use of autologous bone marrow is cheap. 111 addition, harvesting bone marrow is a simple procedure, that does not lead to extra morbidity for the patients [20]. In conclusion, healing of segmental bone defects, treated with granular porous hydroxyapatite, can be improved considerably by the addition of osteoinductive rhOP-1 or osteogenic autologous bone marrow. At 12 wceks, these composite biosynthetic grafts yielded results comparable to those of autologous bone grafting. Therefore, hydroxyapatite combined with rhOP- 1 or bone marrow is a valuable alternative to autograft in the treatment of traumatic bone defects and atrophic nonunions, by which the disadvantages of the harvest of autologous bone can be prevented.
Acknowledgements This work was funded by Merck Biomaterials, Darmstadt, Germany. The authors thank Stryker Biotech, Hopkinton, MA, for providing the OP-1 Device, used in this study. I n addition, the authors thank Connie Niemeyer, M.D., for her contribution in the laboratory work of this study.
References Asahina 1. Sampath TK. Nishimura I. Hauschka PV. Human osteogenic protein-l induces both chondroblastic and osteoblastic differentiation of osteoprogenitor cells derived from newborn rat calvaria. J Cell Biol 1993;123:921-33. Bauer HJ. Ba tier M. Dingeldein E. Zusamniensetzung tind Atifbau einer porosen Hydroxylapatitkeramik biologischer Herk u n f t . Osteosynth Int 1994;2:106-10. Blokhuis TJ. Termaat MF, den Boar FC. Patka P, Bakker FC, Haarman HJThM. Properties of calcium phosphate ceramics in rclntion to their in vivo behavior. J Trauma 2000;48: 179--86. Bruder SP. Fox BS. Tissue engineering of bone. Cell based strategies. Clin Orthop 1999:376S:S68-83. Brtider SP, Kraus KH. Goldberg VM. Kadiyala S. The effect of implants loaded with autologous mesenchymal stem cells on the healing of canine segmental bone defects. J Bone Joint Surg [Am] 1998:80:985 -96. Bucholz RW, Carlton A, Holmes R. lnterporous hydroxyapatite 21sii bone graft substitute in tibia1 plateau fractures. Clin Orthop 1989:240:53 62. Chapman MW. Buchholi! R, Cornell C. Treatment of acute fractures with a collagen
527
[8] Cook SD. Bafes GC, Wolfe MW, Sampath TK. Rueger DC. Recombinant human bone morphogenetic protein-7 induces healing in a canine long-bone segmental defect model. Clin Orthop 1994;301:302-12. [9] Cook SD, Baffes GC, Wolfe MW, Sampath TK, Rueger DC, Whitecloud TS. The effect of recombinant human osteogenic protein-I on healing of large segmental bone defects. J Bone Joint Surg [Am] 1994;76:827-38. [lo] Cook SD, Rueger DC. Osteogenic protein-I. Clin Orthop 1996; 324:29 -38. [ I l l Cook SD, Wolfe MW, Salkeld SL. Rueger DC. Efect of recombinant human osteogenic protein-I on healing of segmental defects in non-human primates. J Bone Joint Surg [Am] 1995: 77:734-50. [I21 den Boer FC, Patka P. Bakker FC. Wippermann BW. van Lingen A. Vink GQM, et al. New segmental long bone defect model in sheep: quantitative analysis of healing with dual energy X-ray absorptiometry. J Orthop Res 1999: 17:654-60. [I31 D’Ippolito G, Schiller PC, Ricordi C, Roos BA, Howard GA. Age-related oateogenic potential of mesenchymal stromal stem cells from human vertebral bone marrow. J Bone Miner Res 1999; 14:1115-22. [I41 Geesink RGT, Hoefnagels NHM, Bulstra SK. Osteogenic activity of OP-l bone morphogenetic protein (BMP-7) in a human fibular defect. J Bone Joint Surg [Br] 1999;81:710-8. [IS] Grundel RE, Chapman MW, Yee T, Moore DC. Autogeneic bone marrow and porous biphasic calcium phosphate ceramic for segmental bone defects in the canine ulna. Clin Orthop 1991; 266:244 58. [I61 Hollinger JO. Brekke J. Gruskin E, Lee D. Role of bone substitutes. Clin Orthop 1991;324:55-65. [I 71 Jarcho M. Calcium phosphate ceramics a s hard tissue prosthetics. Clin Orthop 1981:157:259-78. [I81 Johnson KD. Frierson KE. Keller TS, Cook C, Scheinberg R, Zerwekh J , et al. Porous cerami bone graft substitutes in long bone defects: a biomechanical, histological, and radiographic analysis. J Orthop Res 1996;14:351-69. 1191 Kawamura M. Iwata H, Sato K, Miura T. Chondroosteogenetic response to crude bone matrix proteins bound to hydroxyapatite. Clin Orthop 1987:217:281-92. [20] Lane JM, Tomin E, Bostrom MPG. Biosynthetic bone grafting. Clin Orthop 1999:367S:Sl07-17. [21] Moore DC. Chapman MW, Manske D. The evaluation of a biphasic calcium phosphate ceramic for use in grafting long-bone diaphyseal defects. J Orthop Res 1987;5:356-65. [22] Ohgushi H, Goldberg VM. Caplan AI. Repair of bone defects with marrow cells and porous ceramic. Acta Orthop Scand I989:6O:334-9. [23] Ripamonti U, Ma SS, van den Heever B. Reddi AH. Osteogenin, a bone morphogenetic protein. absorbed on porous hydroxyapatite substrata, induces rapid bone differentiation i n calvarial defects of adult primates. Plast Reconstr Surg 1992:90:38293. [24] Sampath TK, Maliakal JC. Hauschka PV, Jones WK, Sasak H, Tucker R F , et al. Recombinant human osteogenic protein-l ( h o p I ) induces new bone formation in vivo with a specific activity comparable with natural bovine osteogenic protein and stimulates proliferation and differentiation in vitro. J Biol Chem 1992; 267:20352--62. [25] Sato T , Kawamura M, Sato K, Iwata H, Miura T. Bone morphogenesis of rabbit bone morphogenetic protein-bound hydroxyapatite-fibrin composite. Clin Orthop 1991;263:25462. [26] Takaoka K. Nakahara H, Yoshikawa H, Masuhara K , Tsuda T, Ono K. Ectopic bone induction on and in porous hydroxyapatite combined with collagen and bone morphogenetic protein. Clin Orthop 1988:234:250-4
528
I;.C. den Borr
rt
01. 1 Journol qf Oriliopuedic Rcwurch 21 (2003 1 521 -528
[27] Werntz JR, Lane JM, Burstein AH, Justin R, Klein R, Tomin E. Qualitative and quantitative analysis of orthotopic bone regeneration by marrow. J Orthop Res 1996;14:85-93. [ZS] Winn SR, Uludag H, Hollinger JO. Carrier sysiems for hone morphogenetic proteins. Clin Orthop 1999;367:95-106. [29] Wippermann BW, Schratt HE, Steeg S, Tscherne H. Komplikationen der spongiosaentnahme am Beckenkamm. Eine retrospektive analyse von 1191 Fallen. Chirurg 1997;68:128691.
[30] Wippermann B, Zwipp H, Schulz J, Tscherne H. Tierexperimen-
telle Untersuchungen zur Einheilung einer mit Knochenmark auymentierten Hydroxylapatitkerainik (HA) in einem Tibiasegmentdefekt. Vergleich von Ergebnissen nach 3 und 6 Monaten Beobachtungszeit. Langenbecks Arch Chir Forum 1994:521-5. [31] Wozney JM, Rosen V. Bone morphogenetic protein and bone morphogenetic protein gene tamily in hone formation and repair. Clin Orthop 1998;346:26-37.
Journal of Orthopaedic Research
Journal 01' Ortliopoedic Research 21 ( 2 0 0 3 ) S21 52X \\!AM,
elsevier coinlloc'itelorthres
Healing of segmental bone defects with granular porous hydroxyapatite augmented with recombinant human osteogenic protein- 1 or autologous bone marrow Frank C. den Boer 'I,*, Burkhard W. Wippermann b, Taco J. Blokhuis ", Peter Patka ", Fred C. Bakker 'I, Henk J.Th.M. Haarman
Abstract
Hydroxyapatite is a synthetic bone graft, which is used for the treatment of bone defects and nonunions. However, it is a rather inert material with no or little intrinsic osteoinductive activity. Recombinant human osteogenic protein- I (rhOP-I) is a very potent biological agent. that enhances osteogenesis during bone repair. Bone niiirrow contains inesenchynial stcm cells. which are capable of new bone formation. Biosynthetic bone grafts wcrc created by the addition of rhOP-I or bone marrow to granular porous hydroxyapatite. The performancc of these grafts was tested i n ;i sheep model and compared to the results of autograft, which is clinically the standard treatment of bonc dcfccts and nonunions. A 3 cni segmental bone dcfcct was made i n the tibia and fixed with an intcrlocking intraiiiedullary nail. There were five treatment groups: i i o implant ( n =: 6 ) . autograft ( 1 7 = 8). liydroxyapatite alone ( n = 8). liydroxyapatite loaded with rhOP-I ( i i = 8). and hydroxyapatite loaded with autologous bone marrow ( n = 8). At 12 weeks, healing of the dcfcct was evaluated with radiographs, a torsional test to F~ilure,and histological examination of longitudinal sections through the defect. Torsional strength and stillness of the healing tibiae were about two to three times higher for autograft and hydroxyapatite plus rhOP-I or bone niiirrow compared to hydroxyapatite alone and empty defects. The mean values of both combination groups werc comparable t o those of autograft. Thew wcre more unions in defects with hydroxyapatite plus rhOP-l than in defects with hydroxyapatite alone. Although the diflcrences were not significant, histological examination revealed that there was inore oftcn bony bridging of the defect in both combination groups and the autograft group than in the group with hydroxyapatite alone. Healing of bone defects. treated with poroiis hydroxyapatite. can be enhanced by thc addition of rhOP-l or autologous bone marrow. The results of these composite biosynthctic grafts are equivalent to those of autograft. 0 2002 O r t h o p x d i c Rescarcli Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction
In trauma surgery, bone loss creates a major technical and biological problem. Fractures in cancellous bone, such as i n the proximal humerus, thc distal radius, or the tibia1 plateau, often lead to impaction of bone and consequently a defect results after rcduction. Otheicauses of traumatic bone loss are fresh highly coniniinu ted fi-rict it res mid ;it 1-0ph ic ti oti u n i oti s in the d i a ph ys i s of long bones. Autologous bone grafting is widely accepted as the gold standard for the treatment of bone .. .. .. .
..
CorrcsponcIing a u t h o r . T ~ I . +3 : I -20-444-0268: fits: +3 I -20-4440274. l~-iii(ri/ i i t / d w r t : fl.aiikden boer.~~~liotrn;iil.coiii (t'.C. dcii Boer.).
dcfccts and nonunions [20]. However, autologous bone grafting has some serious drawbacks, such as a prolonged opcrntion time [6] and donor site morbidity in about 10-30'%, of the cases [7,29]. I n order to avoid the problems associated with autologous bone grafting, there has been a continuous interest in the use of synthetic bone grafts during the past decades. The most frequently used synthetic bone gr a fts are calcium phosphate ceramics, such as hydroxyapatite, which is known for its excellent biocompatibility [3.16,17]. The intcrconnectivc porous structure of many hydroxyapatite based ceramics acts a s a scaffold and allows the ingrowth of newly formed bone into the pores, which is known a s ostcoconduction [5,19,2123.26,30]. In prospective clinical trials. liydroxyapatite
ncc I . i t l All rights i-e\ei-\ecl. 07.X-O2h6/03/$ - see I'roiit matter. (c) 2002 0rtliop;wdic Rtrearch Societ). Publihed bq E l w icr dol: 10. ~ ~ ~ ~ 6 / ~ ~ ~ ~ ~ 6 - ~ ~ ~ ( , ~ ( ~ ~ ~ ) ~ ~ ~ ~ ~ ~ ~ ~ - ~
522
F.C. diw Boer r t al. I Jourrrul of Orihupuedic Restarch 21 (20031 521 528
ceramics have been used successfully in bone defects, caused by fractures [6,7]. However, hydroxyapatite is a rather inert material with no or little intrinsic osteoinductive activity 1171. In order to enhance new bone formation, growth factors or osteogenic cells are added to hydroxyapatite [4,5,20]. Bone morphogenetic proteins (BMPs) are a group of growth and differentiation factors, that play an important role in embryonic morphogenesis of skeletal and nonskeletal tissues and in osteogenic differentiation during bone repair in postnatal life [31]. One of the BMPs is osteogenic protein-1 (OP-1 or BMP-7), which is manufactured with recombinant DNA techniques [lo]. In vitro studies have shown that recombinant human OP-1 (rhOP- 1) stimulates proliferation and osteogenic differentiation of mesenchymal cells [ 1,241. In criticalsized segmental bone defects in rabbits, dogs, monkeys, and humans, rhOP- 1 induced healing within 2-3 months [8,9,11,14]. These data indicate that rhOP-1 has great potential as a therapeutic agent in bone repair. Bone marrow contains mesenchymal stem cells that possess the capacity to differentiate into osteoblasts [4,13]. Bone marrow is generally considered as an important source of cells, responsible for new bone formation during bone repair. It has been demonstrated in animal models, that fresh autologous bone marrow induces healing of segmental bone defects [15,27]. In this study, composite biosynthetic bone grafts were obtained by the addition of rhOP-1 or autologous bone marrow to granular porous hydroxyapatite. These composite grafts were tested in an established segmental bone defect model in sheep [12]. In order to assess the efficacy of the added rhOP-1 or bone marrow, the results were compared to those of pure granular porous hydroxyapatite. In addition, the performance of the two composite biosynthetic grafts was compared to that of autograft, the gold standard in bone grafting.
Materials and methods Surgicul iwocidwes und .I rucl,: tic,.sign
This study was approved by the institutional review boards for animal experiments. Thirty-nine adult female sheep (thirty-one NorthHolland sheep and eight black-Faced sheep) were used. The mean weight ( z t standard deviation) was 54.2 & 7.6 kg. The animal model, consisting of a 3 cm segmental tibial defect fixed with an interlocking intramedullary nail. has been described previously [12]. In short, the surgical technique was as follows. The left tibia of each animal was operated upon. A 3 cm segmental defect, centred between the innermost interlocking holes of the intramedullary nail, was created by performing two osteotomies with an oscillating saw under constant cooling with saline. The periosteum was not resected. A custom-made A 0 tibial nail (Synthes, Bochum, Germany) was used for fixation. The parts of the tibia, proximal and distal to the defect, were separately reamed. Then. the nail was inserted through the tibial head with the knee in a flexed position. Proximal interlocking was performed, using the A 0 proximal aiming device. A custom-made distal aiming device was used for distal interlocking. Proximally and distally, all three holes were used for interlocking. To reduce differences in outcome between
anirnals, because of differences in the number of osteogenic cells present in the bone marrow at the defect site, all bone marrow in the defect and the medullary cavity was removed by copious irrigation and suction both before and after insertion of the nail. According to the method of treatment of the bone defect, there were five experimental groups: ( I ) empty defects (7 animals), (2) 10 ml morseled autologous corticocancellous bone, harvested from the ipsilateral posterior iliac crest (8 animals), (3) 10 ml granular porous hydroxyapatite alone (8 animals), (4) 10 ml granular porous hydroxyapatite mixed with 2.5 mg rhOP-l (8 animals), and (5) 10 ml granular porous hydroxyapatite mixed with 10 ml autologous bone marrow (8 animals). The soft tissues were closed meticulously over the defect with Vicryl, which ensured containment of the graft materials. As an analgesic 0.5 mg buprenorfine was administered intramuscularly twice daily during the first two days. One sheep with an empty defect sustained a deep infection and this animal was excluded from further analysis. which left six animals in this group. The used calcium Dhosuhate ceramic (Endobon'*, Merck Biomaterials, Darmstadt, Germany) is produced from bovine cancellous bone by a sintering process. This is a manufacturing process under high pressure and at a high temperature (1250 "C), by which the crystallinity of the native hydroxyapatite increases and fusion of crystals occurs. thereby increasing the solidity of the material [2]. Endobon" is an inorganic material. consisting mainly of crystalline hydroxyapatite in the form of granules. sized 2.8 5.6 mm. with an interconnecting pore structure identical to the original cancellous bone. The porosity is 30-80'%1(mean 55%) and the pore size ranges from 100 to 1500 pm with a mean of 450 pm. Prior to surgery. the hydroxyapatite granules were sterilized by dryheating at 180 "C for 30 min. The rhOP-l was bound to bovine bone type I collagen carrier. which is a particulate substance with a size of 75-425 pm. In each animal 2.5 mg rhOP-I and 1.0 g collagen carrier (OP-1 Device, Stryker Biotech, Natick, Massachusetts) was reconstituted with sterile saline, mixed with the hydroxyapatite granules, and implanted in the defect. The 2.5 mg rhOP-1 dose was based on studies with segmental bone defects in monkeys and humans. in which the defects were treated with 1-2.5 mg r-hOP-1 [I I , 141. Autologous bone marrow was aspirated from the ipsilateral posterior iliac crest with a Jamshidi needle on a syringe. A total volume of 10 ml was obtained by three punctures and aspirations of about 3 4 ml in different directions at one site. The bone marrow was mixed with the hydroxyapatite granules and allowed to clot. As part of this cooperative research project. the operations and the evaluation of healing of all animals in the hydroxyapatite plus bone marrow group were performed at the Medical School in Hannover, Germany. Because North-Holland sheep were not available in Germany, black-faced sheep were used in this group: in the other groups, North-Holland sheep were used. The mean weight for the animals with hydroxyapatite plus bone marrow was 59.0 kg and comparable to the mean weight in the other groups, which ranged from 50.0 to 55.4 kg. One-way analysis of variance showed no statistical difference in body weight between the treatment groups ( p = 0.14). The operation technique, the operator (principal author), and the analysis techniques of healing were the same in the hydroxyapatite plus bone marrow group and the other groups. Although the radiographic and biomechanical data of the control groups with empty defects and autologous bone have been described previously [12], these data are included for comparison with the hydroxyapatite groups, which is relevant to the research questions. After 12 weeks all sheep were killed, both tibiae were explanted, and dissected free of soft tissues. The nail was removed and immediately thereafter, radiographical and biomechanical analysis was performed. This 12-weeks end-point was chosen, because in large animal segmental long bone defect models, 12 weeks is a usual and suitable interval for the detection of differences in bone repair among different treatments [8,18,30]. 1
.
Ratliogruphic eouluution
Anteroposterior and lateral radiographs of the left tibia were taken immediately after surgery to evaluate the position of the nail, the locking screws, and the graft material. After 12 weeks, contact radiographs were taken of the explanted tibiae, which were stripped of all soft tissues. Healing of the defect was evaluated by two clinicians, who were unaware of the treatment group. The extent of healing was gra-
F.C. den Boer et ul. I Joumml
of Orthopuedic Reseurcli 21 (2003) 521
ded on a scale of 0 4 , indicating the number of sides, relatively to the medullary canal on the radiographs in two directions, with a uniting bone callus. T o prevent the need for frequent sedation, sequential radiographs in the intervening period till 12 weeks were not made.
528
523
in the number of sides with bony bridging, as assessed by histology. were evaluated with the Kruskal-Wallis test. For all statistical tests a p value of less than 0.05 was considered to be significant. All statistical analyses were performed with the software program SPSS for Windows, version 7.5 (SPSS Inc., Chicago, IL. USA).
Bioiiic~cl7uniccrltesting
Immediately after explantation, a torsional test to failure was performed on both tibiae of each animal. The ends of the experimental tibia were embedded in dental plaster (Vel-mix Stone, Kerr GmbH, Karlsruhe. Germany) in such a way, that the nonembedded portion comprised the 3 cm defect and 1 cm of the diaphysk proximal and distal to it. Because all holes of the locking bolts were embedded, failure through these holes during the test was prevented and the mechanical properties of the defect only determined the outcome of the test. The nonembedded portion in the contralateral intact tibia consisted of an identical 5 cm segment of the shaft, located at the same level. With a custom-built mechanical testing machine the bones were tested to failure by exorotating the proximal end at 20" per minute. Torque was continuously measured at the fixed distal bone end. Torque angular displacement curves were generated, from which the torsional strength (torque at failure) and stiffness were calculated. T o reduce errors associated with geometric or size differences among animals in different groups, torsional strength and stiffness were expressed as a percentage of the value of the contralateral intact tibia. A nonunion was defined as a bone defect with a torsional strength or stiffness, or both. of less than 20% that of the contralateral intact bone [12]. Below the 20'%1 level, the torque^ angular displacement curves showed a slow nonlinear increase in torque indicating plastic deformation. which is consistent with the presence of a fibrous nonunion. N o other criteria, such as the radiographic appearance, were used for essment of nonunions. As shown previously, radiographic evaluation is unreliable for the assessment of nonunions, because new bone formation in the defect can sometimes give a false impression of a uniting bone callus, probably as a consequence of overprojection [12]. The torsional test results were not available for three animals; one in the autogi-aft group due to failure of the recording during the test and two in the hydroxyapatite plus bone marrow group due to an error in the processing of the bones after explantation. IIi.stologiccil tmuljsis
The specimens, containing the 3 cm segmental defect and I cm proximal and distal cortical bone. adjacent to the defect, were fixed in alcohol 70'%,.Longitudinal sections with a thickness of about 2 mm in the central portion of the bone in the frontal plane were sawn with use of a diamond-impregnated band saw (Exakt Apparatebau, Norderstedt, Germany). These undecalcified sections were dehydrated in ascending grades of alcohol and embedded in Technovit (Technovit 7200 VLC, Kulzer GnibH. Wehrheim, Germany). Then 50 pm sections were made, using a sawing and grinding technique (Exakt Apparatebau, Hamburg, Germany). Of each animal one section was stained with Goldner's trichrome stain and one with toluidine blue 0.2'%1.The sections were examined with use of a light microscope. The amount of fibrous tissue and new bone formation and the presence of remodeling were qualitatively assessed. The presence of new lamellar bone and of osteoclasts along the trabeculae of the newly formed woven bone were considered a s signs of remodeling. Bony bridging was defined as a continuous connection with newly formed bone between the proximal and distal end of the defect. Bony bridging was scored as none, one side (medial or lateral), or two sides (medial and lateral), relatively to the medullary canal. Sturitr ic.u
One-way analysis of variance was used to test for differences in the biornechanical data among different treatments, followed by pairwise comparisons between the groups with the post-hoc Duncan test. The numbers of nonunions in the experimental groups were compared with Fisher's exact test and to minimize the influence of multiple comparisons on the results, only a limited number of comparisons were made: hydroxyapatite alone versus empty defects and both combination groups versus hydroxyapatite alone or autograft. Differences among the treatment groups in the median radiographic score of healing and
Results Radiography On the twelve-week radiographs, there was more new bone formation in both combination groups and the autograft group compared to hydroxyapatite alone (Fig. 1). The median radiographic score of healing at this time was 1.5 for empty defects, 2.5 for autograft, 1.0 for hydroxyapatite alone, 2.0 for hydroxyapatite plus rhOP1, and 3.0 for hydroxyapatite plus bone marrow. With the numbers of animals available, there were no significant differences in the radiographic score among the groups 0, = 0.28). Biornechanical tests Torsional strength was comparable for empty defects and hydroxyapatite alone (Fig. 2A). For hydroxyapatite plus rhOP-1, hydroxyapatite plus bone marrow, and autograft, torsional strength was about two to three times higher compared to hydroxyapatite alone or empty defects. There was no difference between the autograft group and both combination groups. For torsional stiffness, the results were comparable (Fig. 2B). According to the biomechanical criterium, there were more nonunions in the empty defect group and the hydroxyapatite alone group compared to autograft, hydroxyapatite plus rhOP-1, and hydroxyapatite plus bone marrow (Table 1). The difference between hydroxyapatite alone and hydroxyapatite plus rhOP-1 was significant ( p = 0.03). Histology In general, the histological findings were in accordance with the results of the radiographs and the biomechanical tests. In all animals, the defect was filled with newly formed woven bone and fibrous tissue. The newly formed bone in the groups with hydroxyapatite was in direct contact with the ceramic, both around the granules and inside the interconnecting pores of the granules, indicating pore ingrowth. By this process of perigranular and intragranular bone ingrowth, the hydroxyapatite granules became incorporated in the newly formed bone. Remodeling of the newly formed woven bone into lamellar bone was present in all cases. Qualitative judgment of the histological sections revealed that in defects treated with hydroxyapatite plus rhOP- 1, hydroxyapatite plus bone marrow, or autograft, the
524
Fig. 1. Anteroposterior and 1;itersl contact radiographs altcr reiiioviil 01' the soft tissues and the nail tit I2 weeks. ( A ) Hydroxyapatitc alone. There ia no new bone formation. the hydroxyapatite granules are still clcai-ly cisible. and the radiographic scoi-c is 7ero. The biomechanical test results are conaistent with ii nonunion. ( B ) Hydroxyapatite plus rhOP- I . As a result ol' abundant new bone formation, the hydroxyapatite granules have become blurred. There is II uniting bone callus at all sidcs around the niedullary canal and the radiographic scow is lour. ( C ) Hydroxyapatite plus autologous bone marrow. Again. there is abundonl new bone Ibrmation with a uniting hone callus a t all four sides and B radiographic score of four.
6o
forsional strength (%)
(A)
.
torsional stiffness (%)
**
***
50 40 -
30
-
70r 60 50
(6)
i
T
~
40 -
20 10
-
10
-
0
Fig. 2. Torsional strength ( A ) and torsional stiflness ( B ) a t 12 weeks. The reaulls tire shown a s a percentage o l ihc value 01' the intact contralateral tibia. The bars indicate the mean values, and the I-bars indicate the standard errors of the mean. ABG: autologous bone grafting, HA: hydroxyapatite. rhOP-I: recombinant human osteogenic protein-I. and BM: bone marrow. (A) Analysis of variance: 11 = 0.02. Post-hoc Duncan test: (*) p < 0.05 vs. HA: ( I * ) p <: 0.05 vs. HA and p < 0.05 vs. empty; (***I 11 < 0.05 vs. HA. ( B ) Analysis o f variance: p = 0.02. Post-hoc Duncan test: (*) p < 0.05 vs. HA iind / J < 0.05 vs. empty.
amount of newly formcd bone was larger and fibrous tissue formation was less than in defects treated with hydroxyapatite alone, (Fig. 3 ) . In some animals with hydroxyapatite alone there was only minimal new bone formation and almost the whole defect consisted of remaining hydroxyapatite granules embedded in fibrous tissue. Correspondingly, bony bridging was observed more often in thc groups with hydroxyapatite plus rhOP-1, hydroxyapatite plus bone marrow, or autograft than in the group with hydroxyapatite alone (Table 2). Howevcr, these difrerences were not significant 0,= 0.25).
Discussion
In traumatic bone dcfects, a bone graft substitute should contain a material with some structural integrity and a space-occupying effect in order to maintain a proper reduction and continuity between the bone fragmcnts during the operation and to enable an adequate osteosynthesis to be performed. It could be argued that thc brittleness of granular porous hydroxyapatite, as used in this study, would make this material unsuitable for this purpose. However, in a traumatic bone defect, a space-occupying effect and little structural
F.C clcw Boer et ul. I Journul of Orthopueiiic Reseurch 21 (2003) 521-528 Table 1 The numbers of unions and nonunions at 12 weeks according to the biomechanical test criteriumd Treatment group
Union
Nonunion
Empty defects Autograft' Hydroxyapatite Hydroxyapatite Hydroxyapatite
2 6 3 8 5
4
+ rhOP-1' + bone marrow'
1 5 0 1
" A nonunion was defined as a bone defect with a torsional strength or stiffness, or both, of less than 20% that of the contralateral intact bone. bThe biomechanical test data were available and thus the presence of a union or nonunion could be assessed for seven of the eight animals in the autograft group and six of the eight animals in the hydroxyapatite plus bone marrow group (see Materials and Methods). rhOP-1 = recombinant human osteogenic protein-I, p = 0.03 compared to hydroxyapatite alone (Fisher's exact test).
support is sufficient, because the major part of the stability is provided by the osteosynthesis and the involved bone will be kept unloaded until healing has been achieved. Although it has been shown that rhOP-1 alone or autologous bone marrow alone are efficient adjuvants in the healing of segmental bone defects [8,9,11,14, 15,271, these materials do not provide structural support. rhOP-1 and bone marrow are (semi-)liquid materials without structural integrity or a space-occupying
525
Table 2 The numbers of animals according to the number of sides wi[h bony bridging, assessed by histological examination of a frontal section, centrally in the defect, at 12 weeks Treatment group
~~
No bridging
One side Two of bridging sides of bridging
2 2 6 2 3
4
~
Empty defects Autograft Hydroxyapatite Hydroxyapatite Hydroxyapatite "rhOP-l
+ rhOP-I" + bone marrow
= recombinant
6 2
6 3
-7
human osteogenic protein-I.
effect. Therefore, no groups with rhOP-1 or bone marrow alone were tested and granular porous hydroxyapatite was taken as the basic material for several bone graft substitutes. In several experimental studies, the combination of hydroxyapatite and BMPs has been tested [ 19,23,25,26]. However, ectopic sites, cancellous bone and calvarial defects, or small animal models were used. In this study, the combination of hydroxyapatite and a BMP was evaluated in a large animal segmental cortical bone defect model. Good results in this model would indicate a highly effective bone graft substitute, which could be useful in many situations of traumatic bone loss
Fig. 3 . Photomicrographs of longitudinal sections of the bone defect in the frontal plane (Goldner's trichrome stain; original magnification. x I ) . The dense tissue at the top and the bottom is the original cortex at the proximal and distal boundary of the defect. The gap between the left and right side is the medullary cavity after removal of the nail. (A) Animal treated with granular porous hydroxyapatite plus rhOP-I, showing abundant new bone formation at both sides. At the lateral (right) side, bone has completely bridged the defect, connecting its cut margins. The score for bony bridging is I . The crack a t the lateral side is caused by the preceding mechanical test. The dark porous particles are the hydroxyapatite granules incorporated in the newly formed bone (arrows). (B) Animal treated with hydroxyapatite alone. There is very little new bone formation at the cut bone ends at the lateral (right) side and in the centre of the defect at the medial (left) side. The rest of the defect is filled with fibrous tissue, surrounding the hydroxyapatite granules (arrows). The score for bony bridging is 0. This animal had a nonunion, which was confirmed in the biomechanical test.
encountered in clinical practise. In addition, the combinations of hydroxyapatite with BMP or bone marrow, both of which were included in the present study, have never been compared to each other. The long bone segmental defect model in the current study is not a consistent nonunion model, since some of the empty defects healed. This can probably be attributed to the retained periosteum, which acts as a source of supply of osteoprogenitor cells [12]. Even in this model, which is not strenuous, different healing responses were found with different graft materials. A disadvantage is that healing was evaluated at only one time-point, i.e. 12 weeks. As a consequence, no conclusions can be drawn about the healing process in the course of the time. For a complete biomechanical and histological evaluation at different times, a lot more animals would be needed for a sufficient number of animals in each group at each time, which is undesirable from an ethical point of view. However, different healing responses with several graft materials, even at one time, indicate which graft materials are most efficient and suitable as bone graft substitutes. In the present study, implantation of pure hydroxyapatite yielded limited new bone formation and the nonunion rate was high and comparable to that of empty defects. In several other animal models with segmental bone defects, the performance of pure hydroxyapatite was comparatively bad [.5,18,21,22]. This is in agreement with the general assumption that hydroxyapatite itself has no osteoinductive activity [17]. Moreover, this ceramic does not contain osteogenic cells. Thus, the bone healing process in defects with pure hydroxyapatite was largely dependent on the osteoconductive properties of the material. However, the granular form of the hydroxyapatite prevented a tight fixation of the ceramic to the adjacent bone, which seems to be a prerequisite for optimal osteoconduction and effective healing [17]. So, the ceramic did not add anything to the regenerative process in the present model. In contrast, efficient healing with much new bone formation and a union rate of loo'% was found, when granular porous hydroxyapatite was combined with rhOP- 1, Calcium phosphate ceramics are suitable carrier materials for BMPs [ 161 and several animal experiments dealing with this topic have been reported. In rodents, purified BMP combined with porous hydroxyapatite induced bone formation in ectopic sites, while negative controls with hydroxyapatite alone did not [19,26]. In cancellous bone defects in rabbits and in calvarial defects in baboons, the addition of purified BMP to hydroxyapatite resulted in significantly more new bone formation [ 193.251. A major disadvantage of purified BMP is the presence of contaminating proteins, that could incite immunogenic responses and carry the risk of disease transmission [X,9, I I , 191. These problems are
prevented by use of rhBMPs, such as rhOP-I, which was used in this study. Similar to almost all previous experimental work with rhOP1, bovine bone type I collagen was used as a carrier material. There are no exact data available on the pharmacokinetics of rhOP- 1 released from different carrier systems. For rhBMP-2, it has been shown that this protein exhibits a sustained release from bovine type 1 collagen and that there is a strong binding affinity to hydroxyapatite mineral with little release after ectopic implantation in rats [28]. Because different BMPs share common structural features [10,31], it could be that rhOP-l in the current model was gradually released from the collagen particles and subsequently retained at the defect site by the hydroxyapatite particles, enabling the protein to exert its osteoinductive effect. Regardless of the exact local pharmacokinetics, the release pattern of rhOP-1 from the collagen particles and the hydroxyapatite granules was consistent with the requirements of the healing bone, as became evident by efficient healing of the defects in this group. It could be argued that the efficacy of rhOP-1 itself has not been proven, because the control group with hydroxyapatite alone did not contain the collagen carrier. However, in experimental segmental bone defects in rabbits, monkeys and humans, the results of the controls with bovine bone type I collagen carrier alone were similar to those of empty control defects with a fibrous nonunion in all cases [9,11.14]. Therefore, it is unlikely that the collagen carrier had an important influence on the bone healing process in the present model. Comparable to the effect of added rhOP-I, the addition of fresh autologous bone marrow to hydroxyapatite resulted in an enhancement of bone healing, which was equivalent to healing with autograft. In the first place, the favourable effect of bone marrow can be attributed to the supply of osteogenic mesenchymal stem cells [4,13,20]. Secondly, the clot associated with bone marrow contains growth factors and the bone marrow cells themselves secrete inductive factors, which stimulate osteogenesis [.5,22]. In other studies with segmental bone defects, an improvement of healing with the combination of hydroxyapatite and bone marrow compared to hydroxyapatite alone was found as well [18,22]. In a 2 cm segmental defect in the sheep tibia, the biomechanical test results were similar for autograft and porous hydroxyapatite plus bone marrow at three and six months [30].In a large prospective randomized clinical trial with fresh traumatic bone defects, the efficacy of the combination of a hydroxyapatite based ceramic and autologous bone marrow has also been demonstrated with results comparable to those of autograft [7]. The results of the present study indicate that rhOP-1 and autologous bone marrow are equally effective as an augmentation of porous hydroxyapatite. rhOP-1 is always off-the-shelf available and can easily be mixed with
I? C. h i Borr
(,I
(11. I Jourriul
of Ortliopurrlic Resrur-cli 21 (2003) 52I-528
granular porous hydroxyapatite; however, the application of recombinant growth factors is rather expensive. Bone marrow needs to be harvested intraoperatively and it is not always possible to obtain a sufficient amount of osteogenic cells. The number of osteoprogenitor cells in bone marrow is relatively small [4,20] and declines with aging [ 131. Conceivably, it is sometimes impossible to harvest a sample of bone marrow with sufficient osteogenic activity in old osteoporotic patients. However, the use of autologous bone marrow is cheap. 111 addition, harvesting bone marrow is a simple procedure, that does not lead to extra morbidity for the patients [20]. In conclusion, healing of segmental bone defects, treated with granular porous hydroxyapatite, can be improved considerably by the addition of osteoinductive rhOP-1 or osteogenic autologous bone marrow. At 12 wceks, these composite biosynthetic grafts yielded results comparable to those of autologous bone grafting. Therefore, hydroxyapatite combined with rhOP- 1 or bone marrow is a valuable alternative to autograft in the treatment of traumatic bone defects and atrophic nonunions, by which the disadvantages of the harvest of autologous bone can be prevented.
Acknowledgements This work was funded by Merck Biomaterials, Darmstadt, Germany. The authors thank Stryker Biotech, Hopkinton, MA, for providing the OP-1 Device, used in this study. I n addition, the authors thank Connie Niemeyer, M.D., for her contribution in the laboratory work of this study.
References Asahina 1. Sampath TK. Nishimura I. Hauschka PV. Human osteogenic protein-l induces both chondroblastic and osteoblastic differentiation of osteoprogenitor cells derived from newborn rat calvaria. J Cell Biol 1993;123:921-33. Bauer HJ. Ba tier M. Dingeldein E. Zusamniensetzung tind Atifbau einer porosen Hydroxylapatitkeramik biologischer Herk u n f t . Osteosynth Int 1994;2:106-10. Blokhuis TJ. Termaat MF, den Boar FC. Patka P, Bakker FC, Haarman HJThM. Properties of calcium phosphate ceramics in rclntion to their in vivo behavior. J Trauma 2000;48: 179--86. Bruder SP. Fox BS. Tissue engineering of bone. Cell based strategies. Clin Orthop 1999:376S:S68-83. Brtider SP, Kraus KH. Goldberg VM. Kadiyala S. The effect of implants loaded with autologous mesenchymal stem cells on the healing of canine segmental bone defects. J Bone Joint Surg [Am] 1998:80:985 -96. Bucholz RW, Carlton A, Holmes R. lnterporous hydroxyapatite 21sii bone graft substitute in tibia1 plateau fractures. Clin Orthop 1989:240:53 62. Chapman MW. Buchholi! R, Cornell C. Treatment of acute fractures with a collagen
527
[8] Cook SD. Bafes GC, Wolfe MW, Sampath TK. Rueger DC. Recombinant human bone morphogenetic protein-7 induces healing in a canine long-bone segmental defect model. Clin Orthop 1994;301:302-12. [9] Cook SD, Baffes GC, Wolfe MW, Sampath TK, Rueger DC, Whitecloud TS. The effect of recombinant human osteogenic protein-I on healing of large segmental bone defects. J Bone Joint Surg [Am] 1994;76:827-38. [lo] Cook SD, Rueger DC. Osteogenic protein-I. Clin Orthop 1996; 324:29 -38. [ I l l Cook SD, Wolfe MW, Salkeld SL. Rueger DC. Efect of recombinant human osteogenic protein-I on healing of segmental defects in non-human primates. J Bone Joint Surg [Am] 1995: 77:734-50. [I21 den Boer FC, Patka P. Bakker FC. Wippermann BW. van Lingen A. Vink GQM, et al. New segmental long bone defect model in sheep: quantitative analysis of healing with dual energy X-ray absorptiometry. J Orthop Res 1999: 17:654-60. [I31 D’Ippolito G, Schiller PC, Ricordi C, Roos BA, Howard GA. Age-related oateogenic potential of mesenchymal stromal stem cells from human vertebral bone marrow. J Bone Miner Res 1999; 14:1115-22. [I41 Geesink RGT, Hoefnagels NHM, Bulstra SK. Osteogenic activity of OP-l bone morphogenetic protein (BMP-7) in a human fibular defect. J Bone Joint Surg [Br] 1999;81:710-8. [IS] Grundel RE, Chapman MW, Yee T, Moore DC. Autogeneic bone marrow and porous biphasic calcium phosphate ceramic for segmental bone defects in the canine ulna. Clin Orthop 1991; 266:244 58. [I61 Hollinger JO. Brekke J. Gruskin E, Lee D. Role of bone substitutes. Clin Orthop 1991;324:55-65. [I 71 Jarcho M. Calcium phosphate ceramics a s hard tissue prosthetics. Clin Orthop 1981:157:259-78. [I81 Johnson KD. Frierson KE. Keller TS, Cook C, Scheinberg R, Zerwekh J , et al. Porous cerami bone graft substitutes in long bone defects: a biomechanical, histological, and radiographic analysis. J Orthop Res 1996;14:351-69. 1191 Kawamura M. Iwata H, Sato K, Miura T. Chondroosteogenetic response to crude bone matrix proteins bound to hydroxyapatite. Clin Orthop 1987:217:281-92. [20] Lane JM, Tomin E, Bostrom MPG. Biosynthetic bone grafting. Clin Orthop 1999:367S:Sl07-17. [21] Moore DC. Chapman MW, Manske D. The evaluation of a biphasic calcium phosphate ceramic for use in grafting long-bone diaphyseal defects. J Orthop Res 1987;5:356-65. [22] Ohgushi H, Goldberg VM. Caplan AI. Repair of bone defects with marrow cells and porous ceramic. Acta Orthop Scand I989:6O:334-9. [23] Ripamonti U, Ma SS, van den Heever B. Reddi AH. Osteogenin, a bone morphogenetic protein. absorbed on porous hydroxyapatite substrata, induces rapid bone differentiation i n calvarial defects of adult primates. Plast Reconstr Surg 1992:90:38293. [24] Sampath TK, Maliakal JC. Hauschka PV, Jones WK, Sasak H, Tucker R F , et al. Recombinant human osteogenic protein-l ( h o p I ) induces new bone formation in vivo with a specific activity comparable with natural bovine osteogenic protein and stimulates proliferation and differentiation in vitro. J Biol Chem 1992; 267:20352--62. [25] Sato T , Kawamura M, Sato K, Iwata H, Miura T. Bone morphogenesis of rabbit bone morphogenetic protein-bound hydroxyapatite-fibrin composite. Clin Orthop 1991;263:25462. [26] Takaoka K. Nakahara H, Yoshikawa H, Masuhara K , Tsuda T, Ono K. Ectopic bone induction on and in porous hydroxyapatite combined with collagen and bone morphogenetic protein. Clin Orthop 1988:234:250-4
528
I;.C. den Borr
rt
01. 1 Journol qf Oriliopuedic Rcwurch 21 (2003 1 521 -528
[27] Werntz JR, Lane JM, Burstein AH, Justin R, Klein R, Tomin E. Qualitative and quantitative analysis of orthotopic bone regeneration by marrow. J Orthop Res 1996;14:85-93. [ZS] Winn SR, Uludag H, Hollinger JO. Carrier sysiems for hone morphogenetic proteins. Clin Orthop 1999;367:95-106. [29] Wippermann BW, Schratt HE, Steeg S, Tscherne H. Komplikationen der spongiosaentnahme am Beckenkamm. Eine retrospektive analyse von 1191 Fallen. Chirurg 1997;68:128691.
[30] Wippermann B, Zwipp H, Schulz J, Tscherne H. Tierexperimen-
telle Untersuchungen zur Einheilung einer mit Knochenmark auymentierten Hydroxylapatitkerainik (HA) in einem Tibiasegmentdefekt. Vergleich von Ergebnissen nach 3 und 6 Monaten Beobachtungszeit. Langenbecks Arch Chir Forum 1994:521-5. [31] Wozney JM, Rosen V. Bone morphogenetic protein and bone morphogenetic protein gene tamily in hone formation and repair. Clin Orthop 1998;346:26-37.