Effect of Chemotherapy on Segmental Bone Healing Enhanced by rhBMP-2

Abstract

Segmental bone defects are challenging clinical problems, and current surgical solutions are associated with high complication rates. In oncologic reconstructive surgery, bone healing will occur coincidently with the administration of chemotherapy to treat the underlying disease. Effective methods of graft modification or bone graft alternatives can be of great help clinically. A series of osteoinductive proteins (bone morphogenetic proteins or BMPs) has been described and shown to enhance bone formation in animal models. This study was designed to evaluate the effect of chemotherapy on bone healing enhanced by rhBMP-2. We used a critical-sized bone-defect rabbit model. Histological and radiological analysis showed that chemotherapy affects both the quantity and the quality of the bone enhanced by the addition of rhBMP-2. These results suggest that the effect of chemotherapy on bone formation could be related to inhibition in a specific pathway stimulated by the rhBMP-2.

INTRODUCTION

The loss of bone that follows operative resection of tumors, traumatic segmental bone loss, or developmental bone defects is a challenging problem. Autogenous, vascularized, and allogenic bone grafts, as well as endoprostheses have been demonstrated to be effective as solutions, but morbidity and complications continue to be troublesome and detract from long-term successful outcomes^1,14. For example, cumulative rates of complication in oncology surgery approach fifty percent and include wound necrosis, infection, nonunion, fracture, prosthesis loosening, and immunologic complications^{19,23,25,26,28,30,33}. In addition, chemotherapy, which has been proven to improve the relapse-free survival time of patients with certain primary bone sarcomas, must be initiated early in the treatment course and may be required for as long as a year after surgical resection¹¹. Although chemotherapy is effective in controlling cancer cell growth, it also has systemic effects, especially in the bone marrow. Effective methods of graft modification, or bone graft alternatives to overcome these problems could be of great help clinically.

The formation, maintenance, and regeneration of bone are complex processes involving the interactions of many cellular elements with systemic and local regulators. Recent gains in understanding of the biology of fracture healing and the availability of specific macromolecules have resulted in the development of novel treatments for bone defects. A series of osteoinductive proteins (bone morphogenetic proteins or BMPs) has been described and shown to enhance bone formation in animal models^{1,2,6,9,15,29,31,35}. The major capacity of BMPs is to induce the differentiation of both pre-osteoblastic cells and non-committed mesenchymal cells. In addition, using recombinant molecular techniques, BMPs can be produced in large quantities, thus paving the way for their potential use in the healing of bony segmental defects.

Evaluation of BMPs to date has been limited to the treatment of tibial nonunion, and applicability to other indications awaits further experimental and clinical research. For BMPs to be used in the treatment of musculoskeletal tumors it is imperative that we understand the modifying effects chemotherapy may have on the bone healing induced by BMPs. We developed a model in rabbits to evaluate the effects that chemotherapy has in the healing of critical-sized bone segmental defects¹ treated with insoluble bovine bone carriers added with recombinant human bone morphogenetic protein-2 (rhBMP-2).

MATERIAL AND METHODS

Experimental Design

Unilateral two-centimeter critical-segmental bone defects were created in the radial diaphyses of 45 young adult New Zealand White rabbits. Six groups of animals were studied: Group 1: The defect was left empty (untreated controls); Group 2: Defect filled with a collagen-carrier containing zero micrograms of rhBMP-2; Group 3: Defect filled with a collagen-carrier containing thirty micrograms of rhBMP-2; Groups 4, 5, 6: Same as surgically treated groups one, two and three, but each group received intravenous doxorubicin and cisplatin. This study was approved by our institutional Animal Care and Use Committee.

Operative Procedure

The surgical approach to the radius was identical in all rabbits. All operative procedures were performed in a surgical suite using intravenous anesthesia with Ketamine/ Xylazine/Acepromazine as described.¹ Cephalothin (40 mg/kg), was administered prior to surgery and twice a day for two days postoperatively. A four-centimeter superomedial incision was made and the soft tissues overlying the radial diaphysis were dissected. A two-centimeter bone segment was removed with the use of an oscillating saw and the defect was filled with the experimental delivery system. Muscle, fascia and skin were closed in a standard fashion. The animals were monitored closely for signs of discomfort or surgical complications post-operatively. Analgesics were administered based on observation by a veterinarian as individually needed to insure the animals' comfort. Throughout the experiment, all animals remained individually caged.

Preparation and Placement of the Delivery System Containing rhBMP-2

The experimental delivery system consisted of a carrier of insoluble bovine bone collagen (Helistat, Integra Life Sciences, Painsbore, NJ) reconstituted with zero (Groups two and five) or 30 micrograms (Groups three and six) of recombinant human bone morphogenetic protein-2 (rhBMP-2) (Genetics Institute, Andover, MA). At the time of the operation, the sterilized collagen carrier was loaded with the reconstituted rhBMP-2. After the twenty-millimeter bone segment had been removed, the gap was irrigated with sterile warm normal saline and the delivery system was positioned in the defect. The muscles, augmented by the soft-tissue closure, retained the graft.

Doxorubicin and Cisplatin Treatment

The chemotherapy groups (Groups 4, 5 and 6) received 2.5 milligrams per kilogram of body weight of both doxorubicin and cisplatin intravenously four days before the index operation and again at seven and 14 days after the procedure. Hydration during drug administration was performed to decrease nephrotoxicity.

Radiographic Methods

All the rabbits were radiographed postoperatively and at weekly intervals. To insure proper positioning during radiographs, all rabbits were anesthetized with ketamine/xylazine/acepromazine IM. Antero-posterior and lateral radiographs were taken at weekly intervals to evaluate bone healing. The radiographs were interpreted by three of the investigators who were blind to treatment type. Radiographic evaluation was performed by measuring the area of periosteal callus and diaphyseal bone at the osteotomy site using digitized images of the x-rays. Periosteal callus and diaphyseal bone were outlined along the bone between the two bone ends in both the lateral and antero-posterior films. The area was calculated from each view and is expressed as a ratio of periosteal callus to diaphyseal bone. Image J analysis software (NIH) was used for the analysis.

Histological Methods

After the animals were euthanized, radii specimens were stripped of surrounding soft tissues (except directly over the fracture site), fixed in ten percent neutral buffered formalin and decalcified in four percent formic acid for seven to ten days. Specimens were then embedded in paraffin and sectioned longitudinally (5 thick). Three sections from the middle of the diaphysis were stained with hematoxylin and eosin. Stained sections were photographed and magnified 140 times to create an enlarged print of the fracture. Using Image J analysis software (NIH), the relative proportion of bone fracture callus was determined. Points that fall on cortical bone, artifacts, fibrin clot or blood vessels were subtracted from the total number of points.

RESULTS

Radiographic Analysis

A bone bridge developed at two weeks after the procedure in the groups where we used the insoluble bovine bone collagen carrier. Although we found no statistically significant difference between the average optical densities in the different groups, increase in the bone density and area was evident during the next four weeks. The use of BMP improved both area and density of the new bone formation and the quality of bone: the cortical bone formation and the appearance of a new medullary cavity were more evident in the animals treated with BMP. The addition of chemotherapy resulted in impairment of bone formation, with a decrease in the area and density of new bone (Table 1). In the different groups the findings were:

TABLE 1. Radiographic measurements at eighth week.

Group	Average Initial Defect (mm)	Defect Void	Defect Bridging
Control	21	2 of 9	7 of 9
Control +Chemo	25	4 of 6	5 of 6
Helistat	27	2 of 9	7 of 9
Helistat +Chemo	24	4 of 7	3 of 7
rhBMP2	23	0 of 9	9 of 9
rhBMP2+Chemo	25	3 of 5	2 of 5

Controls:

Although no bone formation was expected from this group, the fact that the rabbits studied were still young adults allowed for some bone formation in the gaps. This probably was due to the remaining periosteum not resected during the surgical procedures (Table 1, Figure 1).

Figure 1. Average Optical Density of Bone.

Average radiographic optical density at eight weeks

Insoluble bovine bone collagen carrier:

Bone formation started at two weeks and progressed through weeks four and six in a uniform fashion. The new bone was fused to the edges of the radius by the end of week two, and there were signs of recanalization of the medullary cavity by the eighth week. Seven of the nine cases bridged the gap and the average optical density was 6.4 (Table 1, Figure 1).

Insoluble bovine bone collagen carrier plus BMP:

Bone formation started earlier in this group and was advanced by the second week, with important bone formation and central recanalization of the medullary cavity by the end of week eight (Figure 1). Cortical bone is apparent in the X ray images at that time Figure 2. The gap was bridged in all the cases, improving the results of the group without BMP addition. The average optical density obtained (6.5) was the highest in all groups (Table 1, Figure 1).

Figure 2.

Radiograph of a rabbit treated with Helistat and BMP at eight weeks. Note the complete fusion at both ends of the created gap with the formation of a totally recanalized bone, with cortical bone formation both in the medial and the lateral cortex.

Insoluble bovine bone collagen carrier plus Chemotherapy:

Bone formation started at two weeks and progressed in amount of bone and density. The bone was fused to the radius extremes by the end of week two and there were signs of recanalization of the medullary cavity by week six. However, the gap had a void in four of seven cases, showing a deleterious effect of the chemotherapy in this area. The average optical density was 5, which was less than the density in the group without chemotherapy (Table 1, Figure 1).

Insoluble bovine bone collagen carrier plus BMP plus chemotherapy:

Bone formation was evident by the second week, with progression in area and density over the next four weeks. There was a tendency toward recanalization of the medullary cavity by the end of eighth week (Figure 3). There was an important decrease in the bridging with a failure to accomplishing it in three out of five cases. The average optical density of 5 was also decreased compared to the group without chemotherapy.

Figure 3.

Radiograph of a rabbit treated with Helistat, BMP and chemotherapy at eight weeks. The bone density is decreased when compared with group 5. Note that the bone formed does not have the new medullary channel formation as noted in the previous groups.

Histological Analysis

Controls:

There was some bone formation in non-treated rabbits, with minimal periosteal bone formation of immature bone. The area of bone formation was 566 pixels. For the control-plus-chemotherapy group the area was 415 pixels using Image J software.

Insoluble bovine bone collagen carrier:

The bone formed in this group was characteristically woven, non-organized bone, with recanalization of the medullary cavity at both extremes in the union with the normal radius. There were no signs of inflammation, fibrous tissue or foreign body reaction in any of the cases. The area of bone formation was 375 pixels using Image J software.

Insoluble bovine bone collagen carrier plus chemotherapy:

Characteristically, these cases formed poor organized woven bone. This osseous tissue filled the osteotomy gap, and there was complete fusion with the cortices at both sides of the radius. There were no signs of inflammation, fibrosis, foreign body reaction or cartilage formation. The area of bone formation was 880 pixels using Image J software.

Insoluble bovine bone collagen carrier plus BMP:

An improvement in bone organization was noted in this group, with complete recanalization of the medullary cavity in two cases and formation of cortical bone bridging the osteotomy. The amount of new bone formation was improved when compared to the non-BMP treated groups, and the bone unions were completely fused at both ends of the radius. In these cases there were no signs of inflammation, foreign body reaction, fibrosis or cartilage formation. The area of bone formation was 1110 pixels using Image J software (Figure 4).

Figure 4.

Histological image of the bone formed after eight weeks in an animal treated with Helistat and BMP. Note lamellar organization of the bone in the cortices and the formation of a medullary cavity. There is no evidence of foreign body reaction or fibrous tissue formation.

Insoluble bovine bone collagen carrier plus BMP plus chemotherapy:

There was a decrease in the quality of the bone formed in these cases with less bone area, more woven bone appearance, and less cortical bone formation. However, there was full bone fusion at the ends of the radius. There were no signs of inflammation, foreign body reaction, infection or formation of fibrous or cartilaginous tissue. The area of bone formation was 940 pixels using Image J software (Figure 5).

Figure 5.

Histological image of the bone formed after eight weeks in an animal treated with Helistat, BMP and chemotherapy. Note the woven orientation of the bone with no cortical organization. No signs of inflammation or foreign body reaction.

DISCUSSION

Limb reconstruction after tumor resection continues to be a major challenge in orthopaedic oncology. Many techniques are available, but the most appropriate choice is dictated by local tumor factors (size, location, stage, etc.) and patient factors (age, activity level, systemic disease, etc.). The skeletally immature patient presents special problems, including increased demands on the reconstructed limb, risk for growth disturbances, and the need for long-term optimal results.¹⁴ Factors considered to affect bone healing include inadequate soft-tissue coverage, need for multiple surgeries, and adjuvant chemotherapy or radiation therapy^16,30. The search for an acceptable substitute for autogenous and allograft bone has involved proteins that induce bone formation in vivo.

The molecular cloning of the bone morphogenetic proteins and their subsequent expression in recombinant systems has permitted the use of these molecules in a variety of experimental models^{1,2,6,15,29,31,35}. Fifteen BMPs have been characterized and cloned so far. The BMPs are multifunctional proteins and have various effects on cell growth and differentiation according to dosage and cell type. The major characteristic of BMPs is their capacity to induce differentiation of both pre-osteoblasts and non-committed mesenchymal cells. Unlike tissue growth factor beta (TGFb), this potential to commit mesenchymal cells to differentiation is specific to BMPs. The effect of BMPs on cell proliferation is variable: Proliferation of osteosarcoma cells is stimulated by BMP-7 and BMP-2, while proliferation of osteoblasts is stimulated by BMP-7, but inhibited by BMP-2 in vitro. The ability of rhBMP-2 to stimulate local bone formation was observed in this study as accelerated callus formation and maturation demonstrated by the histologic and radiographic results. This accelerated bone induction was presumed to be an effect of rhBMP-2's well-documented ability to induce the local differentiation of uncommitted mesenchymal and osteoblast precursor cells into osteoblasts^26,19. This study, in accordance with previous reports,^1,2 shows an improvement in bone healing in a model of critical-sized bone defects treated with bone collagen carriers and rhBMP-2 compared to normal controls.

Importantly, almost all patients with a high-grade bone sarcoma will have adjuvant chemotherapy that must be initiated early in the reconstructive plan and may be required for as long as a year after surgery.^10,11 Many chemotherapeutic drugs used in adjuvant tumor treatment are known to exert their effects on rapidly proliferating cells. Standard doses of many chemotherapeutic agents cause temporary bone marrow suppression occurring one to two weeks post-administration. Therefore chemotherapy, when used in combination with limb salvage procedures, may inhibit bone formation. Combination therapy with doxorubicin (Adriamycin) and cisplatin has been found to be an effective treatment of bone sarcomas^3,24. Doxorubicin exerts its cytostatic effect by intercalating between DNA base pairs, thus inhibiting DNA synthesis and DNA-dependent RNA synthesis. Cisplatin is thought to act by producing inter- and intra-strand cross-links of cellular DNA, thus inhibiting transcription. The use of this drug combination in adjuvant treatment of musculoskeletal tumors has resulted in greatly improved results.^10,11,24

The consequences of chemotherapy administration on the process of bone formation remain controversial. Negative effects on bone healing and bone turnover have been found with reduced bone formation, both in normal bone and after fracture, and delay of the incorporation of autografts in segmental cortical defects in animal models. ^{4,7,8,12,13,17,18,20,22,27} Nilsson et al.¹⁸ evaluated the effect of doxorubicin and methotrexate on orthotopic bone and on the induction of experimental heterotopic bone in rats. They found that doxorubicin treatment, at the time of implantation of bone matrix, caused reduced amounts of bone formation (30-35 percent) in heterotopic bone, whereas orthotopic bone was unaffected. However, six weeks after the treatment the net effect on the induced bone decreased. The results suggested that bone formation is sensitive to inhibition by anti-neoplastic agents, especially in conditions in which recruitment of new bone-forming cells is required. Similar conclusions can be drawn from a rat study by Pelker et al.²⁰ who studied doxorubicin and methotrexate in a rat osteotomy model. They found a significant decrease in the torsional strength of healing osteotomies in animals receiving chemotherapy while observing no strength difference between intact bones of treated animals and controls. Khoo³¹ looked at the effects of preoperative doxorubicin on wound and bone healing in rabbit femoral fractures and reported decreases in wound breaking strength and torsional bone strength in animals that received the agent within one week of surgery. Prevot et al.²¹, using lengthening of adult rabbit tibias, found a slight delay in ossification when methotrexate and doxorubicin were used.

Like doxorubicin, cisplatin has also been demonstrated to alter bone and soft tissue healing in animal models. Zart et al.³⁴ studied the effect of cisplatin on syngenic and allogenic cortical bone graft incorporation in rats. This work demonstrated smaller total bone areas in the grafts of cisplatin-treated animals. They also found that revascularization and host-graft union were both slower in cisplatin-treated animals compared to controls. These differences were more pronounced in the animals receiving frozen allografts than in those getting the fresh synthetic grafts. Young et al.³², using diaphyseal segmental replacement in dogs, observed that cisplatin in the postoperative period caused a delay in extra-cortical formation and significantly reduced both graft resorption and new bone formation.

In this study we found that chemotherapy affects both the quantity and the quality of the bone enhanced by the addition of rhBMP-2 to a collagen matrix. One possible mechanism would be an increase in the chemotherapy related apoptosis of dividing cells previously stimulated by BMP since the therapeutic agents used in this study act in actively dividing cells. Also, since BMP stimulates cell division and differentiation by activating the cells' DNA machinery, this could create an increased number of cells that chemotherapy could target.