Vol 8, No 1, January/February 2000 3 Of the various bone-graft materials available, autologous bone graft is the standard for arthrodesis of the spine. The relative superiority of autologous bone grafting is due to its osteoinductive, as well as its osteoconductive and osteogenic, properties. Autologous bone graft is thought to contain growth fac- tors such as bone morphogenetic proteins (BMPs). 1 These proteins have been shown to induce bone formation through endochondral mechanisms, leading to their possi- ble use in isolated form to achieve spinal fusion and to repair long- bone fractures. 2,3 In 1979, Urist et al 4 first showed that proteins with bone morpho- genetic properties could be extracted from animal cortical bone by diges- tion of the demineralized cortical bone matrix with bacterial collage- nase and solubilization of the digest in a neutral mixture of salt and ethylene glycol. The BMP ex- tracted in this manner was found not to be species-specific; that is, BMP extracted from a rabbit in- duced new bone in rats and there- fore might do so in man as well. Bauer and Urist 5 then isolated human BMP (by using a 4M guani- dine hydrochloride solution) that was capable of inducing bone for- mation in the thigh muscles of athymic nude mice. Since that time, more BMP types, better isolation techniques, and the advent of recombinant cloning techniques have made the use of BMP in the clinical setting a reality. Currently, the use of BMP in humans is restricted to trials evaluating its use in achieving spinal arthrodesis 6 and in treating nonunions and other difficult problems. Possible future uses for BMPs include enhancement or re- placement of autologous bone graft, treatment of delayed union or nonunion, compensation for patient factors such as nicotine use, facilitation of spinal or recon- structive arthrodeses, supplemen- tation of biologic ingrowth, and management of osteonecrosis and certain pathologic osteopenias. 7 Additionally, current basic science research is evaluating the efficacy in animal models of a single dose of recombinant human BMP-2 (rhBMP-2) versus the use of mar- row cells genetically transformed to produce BMP-2. Dr. Zlotolow is Clinical Research Fellow, Department of Orthopaedic Surgery, The Rothman Institute at Thomas Jefferson University, Philadelphia. Dr. Vaccaro is Associate Professor of Orthopaedic Surgery, The Rothman Institute. Mr. Salamon is Research Assistant, Thomas Jefferson University. Dr. Albert is Associate Professor of Orthopaedic Surgery, The Rothman Institute. Reprint requests: Dr. Vaccaro, The Rothman Institute, 925 Chestnut Street, Philadelphia, PA 19107. Copyright 2000 by the American Academy of Orthopaedic Surgeons. Abstract The attainment of a stable arthrodesis is critical to the successful management of some types of spinal disorders. Autologous iliac-crest bone graft has been the most commonly utilized substance associated with predictable healing in spinal fusion applications. Although alternative graft substances exist, these have not been shown to be as uniformly effective in achieving spinal fusion. Because of the morbidity associated with bone autograft harvest, there is increasing inter- est in alternative graft substances and especially in the osteoinductive abilities of bone morphogenetic proteins (BMPs). Several animal models have demon- strated that BMP-containing allograft or synthetic carrier medium is as effective as or superior to autograft bone in promoting spinal fusion. Furthermore, the limited number of human trials utilizing BMPs to treat nonunions in the appendicular skeleton indicate that the results found in animal models will be reproducible in the clinical setting. J Am Acad Orthop Surg 2000;8:3-9 The Role of Human Bone Morphogenetic Proteins in Spinal Fusion Dan A. Zlotolow, MD, Alexander R. Vaccaro, MD, Michael L. Salamon, and Todd J. Albert, MD Perspectives on Modern Orthopaedics Classification The BMPs, with the exception of BMP-1, are grouped as a family within the transforming growth factor β (TGF-β) superfamily of dimeric, disulfide cross-linked growth and differentiation factors (Table 1). Although BMPs have been studied primarily for their osteoinductive properties, they may also be found in extraskeletal tissues, where they function to reg- ulate the development of other organ systems. 8 The sole criterion for BMP classification is the induc- tion of ectopic bone formation in a standard in vivo rodent assay sys- tem. Individual BMPs are grouped on the basis of their amino acid sequences and molecular structural components. The BMP family encompasses more than 12 proteins, 9 of which have demonstrated an ability to induce ectopic bone formation in an in vivo assay system. 9 The BMPs used for the initial basic and clinical research endeavors were originally extracted from deminer- alized human cortical matrix. A number of extraction schemes have been developed, but all are com- plex and do not reliably yield usable quantities of BMP, as only 1 to 2 µg of BMPs are present in a kilogram of cortical bone. Further- more, the presence of contaminants in these extracts is of some concern in this era of blood-transmissible infections. 7 The application of recombinant DNA technology in the manufac- turing of genetically engineered BMPs as early as 1988 2 has allowed for a virtually limitless supply of a few rhBMPs (e.g., rhBMP-2, -4, and -6,Genetic Institute, Cambridge, Mass; OP-1 rhBMP-7, Stryker Bio- tech, Natick, Mass) for basic re- search and clinical trial applica- tions. Despite this, the mechanism of action and the role of each of the seemingly redundant BMPs remain controversial. What little is known about these proteins has been stud- ied mostly in in vitro assay systems and in clinical trials in animals. Osteoinductive Properties Analogous to the cascade theory of bone formation witnessed in em- bryonic endochondral ossification and fracture callus formation, BMPs induce bone formation in a stepwise fashion. The sequence includes chemotaxis of progenitor cells, proliferation of mesenchymal cells, differentiation of chondro- cytes, calcification of cartilage matrix, angiogenesis and vascular invasion, bone differentiation and mineralization, and bone remodel- ing and marrow differentiation. Once formed, the bone seems to function as typical endochondral bone, responsive to both internal and external stimuli. 8 The BMPs may themselves act in a stepwise fashion. The presence of one BMP may induce the expres- sion of other BMPs. Moreover, two different BMPs may join to form disulfide-linked heterodimers, and these heterodimers may bind and activate BMP receptors with greater affinity than that of BMP homo- dimers. The combinations of BMP-2 with BMP-7 and BMP-2 with BMP-6 have been shown to be five- to ten- fold more potent in inducing carti- lage and bone formation than BMP-2 alone. 10 This, along with the dis- covery of a joint BMP-2ÐBMP-4 re- ceptor, suggests that BMPs may function as heterodimers in vivo. 11 Alternatively, BMP-2 and BMP-4 may induce endochondral bone for- mation, while BMP-6 may preferen- tially induce direct membranous bone formation. There is evidence that BMP-7 stimulates mRNA levels of BMP-6 while decreasing mRNA levels of BMP-2 and BMP-4, 12 and that BMP-2 stimulates BMP-3 and BMP-4 mRNA levels. 13 Boden et al 14 found that physiologic levels of the glucocorticoid triamcinolone ace- tonide promoted osteoblast differen- tiation from fetal calvarial cells by inducing BMP-6 production. The effects of the glucocorticoid could be blocked by BMP-6 antisense DNA (antisense DNA blocks sister-strand mRNA translation), indicating that BMP-6 mRNA translation to protein is a necessary step downstream of glucocorticoid production in the osteoinductive pathway. Further- more, BMP-6 was found to be 2 to 2.5 times more potent than either BMP-2 or BMP-4, and to not be as potentiated by a glucocorticoid infu- sion as the other two. 15 The cascade of bone osteoinduc- tion may be initiated by a BMP, leading to the controlled expres- sion of other BMPs, which may then work synergistically to stimu- late bone formation. In sponta- neous bone formation, the gluco- corticoid may be the principal or early signaling molecule, and may initiate a signal amplification cas- cade beginning with BMP-6. Due to its function early in the cascade, BMP-6 appears to be the ideal osteoinductive protein for investi- gation in clinical trials. Segmental-Defect Models The potential of BMPs was initially investigated by utilizing segmental- defect animal models. In 1982, Takagi and Urist 16 showed that extracted bovine BMP could be Human Bone Morphogenetic Proteins in Spinal Fusion Journal of the American Academy of Orthopaedic Surgeons 4 Table 1 Family Groups Within the TGF-β Superfamily TGF-β family Inhibin/activin family MŸllerian inhibiting substance family Decapentaplegic BMP family used to heal large femoral diaphy- seal defects in rats. The authors used an omega pin to distract the two ends of the femur and to block the migration of osteoinductive and osteoconductive elements within the marrow. Although a variety of graft materials were evaluated, including variable doses of BMP without a carrier, only the defects treated with BMP and autologous marrow healed 100% of the time. The earliest experimental studies of rhBMP-2 also involved iatrogeni- cally produced bone defects in an animal model. Yasko et al 17 found that rhBMP-2 combined with bone marrow in a rat segmental femoral- defect model produced union at a rate of 100%, three times superior to the rates achieved with autogenous cancellous bone graft. Similar bone-defect studies have been con- ducted in rabbit tibia and ulna, sheep femur, and canine spine and mandible. Toriumi et al 18 showed that rhBMP-2 could effectively in- duce bone formation in the man- dible, which embryologically follows the intramembranous pathway, un- like previous studies that focused on endochondrally derived bones. Early results in human clinical trials for tibial nonunions suggest out- comes equivalent to those obtained with use of autologous bone graft. Posterior Lumbar Spine Fusion Posterior lumbar fusions in humans, unlike anterior fusions, may not be loaded in compression and therefore occur less predictably. Clinically im- proving the rate of posterior fusion may require use of an osteoinduc- tive substance such as autograft or BMP. At the present time, the only human BMPs tested in clinical ani- mal spine fusion models have been rhBMP-2 and rhBMP-7. In addition to investigating the efficacy of BMPs, studies of these substances have pro- vided information on the optimal delivery system or carrier, potential complications, and the effects of dosage manipulation on fusion rate and success. Early Results As early as 1989, Lovell et al 19 evaluated the effect of extracted BMP on experimental posterior inter- vertebral spine fusions in mature mongrel dogs. Radiologic and histo- morphometric evaluation showed that the fusion levels augmented with BMP had two to three times more new-bone formation than con- trol levels. Fusion occurred at 71% of the levels treated with the BMP but only 14% of the control levels. How- ever, the polylactic-acid polymer car- rier utilized was not resorbed com- pletely, suggesting that a better carrier material needs to be found. Schimandle et al 20 demonstrated higher fusion tensile strength and stiffness in rabbit posterior lumbar- spine fusions supplemented with rhBMP-2 as compared with those in animals that received autogenous iliac-crest bone graft. All animals treated with rhBMP-2 demonstrated solid fusion, as evaluated by manual palpation and radiographic exami- nation, compared with only 42% of the autograft animals. In another study, performed in a posterolateral intertransverse-process model in rabbits, rhBMP-2 was shown to reverse the inhibitory effects of the nonsteroidal anti-inflammatory drug ketorolac on fusion rate. 21 Dose Response Using rhBMP-2 with a collagen carrier in a canine spinal fusion model, David et al 22 showed a dose- dependent osteoinductive effect, with a 100% clinical and radio- graphic fusion rate. Using a canine fusion model, Sandhu et al 23 dem- onstrated that rhBMP-2, delivered at a dose of 2,300 µg in a porous polylactic-acid polymer, was supe- rior to autologous iliac-crest bone graft in achieving a single-level lumbar arthrodesis. In a later study, this group investigated the dosage level at which no further osteoinduction was achieved. Re- combinant human BMP-2 was implanted at multiple doses, in- cluding 58, 115, 230, 460, and 920 µg. Histologically, abundant bone formation occurred in all specimens containing rhBMP-2 by 3 months. 24 The data from this study and the one using 2,300 µg of rhBMP-2 showed no mechanical, radiographic, or histologic variations in the quality of intertransverse-process fusion re- sulting from a 40-fold increase in rhBMP-2 dosage (58 µg to 2,300 µg). The rhBMP-2 dosage require- ment for successful fusion was also investigated by Boden et al 25 in a primate model of laparoscopic anterior lumbar interbody arthro- desis in which a titanium-threaded interbody fusion cage was used. Before insertion, the cages were soaked in either high-dose (1,500 µg/mL) or low-dose (750 µg/mL) rhBMP-2. Although bone formation and spinal fusion were achieved at both doses tested, the higher dose produced a more rapid fusion re- sponse. In humans, the optimal dosage based on the optimal carrier remains to be determined, although dose-response studies have been carried out in nonspinal locations in humans. Carrier Medium The role of the carrier medium is to allow the BMP molecules to be applied in an easily reproducible localized fashion and to prevent rapid uncontrolled diffusion into the surrounding tissues. The ideal car- rier medium would be biocompatible, completely resorbable, structurally stable, and easy to manipulate. A variety of media, including biologic substances, polymers of different types, and even titanium sponges, have been investigated in the ap- pendicular skeleton. Dan A. Zlotolow, MD, et al Vol 8, No 1, January/February 2000 5 Sheehan et al 26 examined various carrier media in posterior lumbar spine fusions in an adult female ca- nine model. Using the same surgi- cal approach, the authors compared four different treatments: autoge- nous iliac-crest bone alone, bovine type I collagen gel and autogenous iliac-crest bone, type I collagen gel combined with autogenous iliac- crest bone and rhBMP-2, and con- trol (sham) without an implant. There was considerably more bone formation at the sites containing rhBMP-2 than at the sites contain- ing autogenous iliac-crest bone graft either alone or combined with the collagen gel carrier. Biome- chanical testing of the explants demonstrated superior strength of the rhBMP-2 fusion sites. Fischgrund et al 27 evaluated the augmentation of autograft by using rhBMP-2 combined with various car- rier media in a canine lumbar spine fusion model. The carrier media included a collagen ÒsandwichÓ made of collagen sheets, collagen ÒmorselsÓ (collagen sheets cut into small pieces), open-pore polylactic acid, and a polylactic acidÐglycolic acid sponge sandwich, with auto- graft alone or in combination with rhBMP-2 as controls. Greater in- creases in bone fusion mass were recorded at all levels that involved rhBMP-2 as compared with levels containing autograft alone. In addi- tion, carriers that were combined with morselized bone graft offered easier technical handling and appli- cation during the operative proce- dure. Use of the polylactic acidÐ glycolic acid sponge sandwich as a carrier was associated with a greater incidence of voids within the fusion mass compared with the use of colla- gen sandwich, collagen morsels, open-pore polylactic acid, or auto- graft with rhBMP-2 alone. Ulti- mately, no important difference in the efficacy of the various carrier media could be determined from this study. However, the addition of rhBMP-2 to autograft enhanced the volume and maturity of the resulting fusion mass. The most recent studies suggest that another approach to the use of BMP-2 in spinal fusions may be pos- sible. Rather than utilizing a single dose at the time of surgery, geneti- cally transformed marrow cells were implanted, which produced continuously generated quantities of BMP-2. Boden et al 28 transplanted marrow cells in rats with cDNA for an osteoinductive protein and had a 100% fusion rate, compared with 0% for controls. Similar results were observed with adenoviral vectorÐ transformed, BMP-2Ðproducing marrow cells in a rat fusion model; the effects were comparable to those obtained with implanted rhBMP-2. 29 Decortication and Minimally Invasive Techniques The value of host-bone decortica- tion was studied by Sandhu et al 30 in a dog intertransverse-process fusion model in which rhBMP-2 was used. The argument for the necessity of decortication to achieve fusion is that it unmasks marrow elements, osteoinductive proteins, inflammatory cells, the local blood supply (including the initial hema- toma), and osteogenic cells at the fusion site. In that study, there was no statistical difference in the clini- cal and radiographic fusion rates between decorticated and non- decorticated fusion sites. Further- more, as the dosage of rhBMP-2 was increased, there was little his- tologic discrimination between fusions in decorticated spines and those in nondecorticated spines. Boden et al 31 demonstrated the feasibility, efficacy, and safety of a minimally invasive application of rhBMP-2 delivered in a collagen sponge carrier in both a rabbit and a nonhuman primate (rhesus mon- key) intertransverse-process model. This technique minimizes the mor- bidity of paraspinal muscle dener- vation and devascularization seen with open intertransverse-process fusion techniques while providing effective fusion. In a subsequent study, Boden et al 32 again utilized the minimally invasive video-assisted lateral inter- transverse-process approach in rhe- sus monkeys with a different carrier. Instead of a collagen carrier, which was criticized for its nonrigidity and its lack of predictability of rhBMP dose delivery, the researchers used a rectangular-block ceramic carrier made of hydroxyapatite (60%) and tricalcium phosphate (40%) with 9 mg of rhBMP-2. At follow-up, the intertransverse processes in all five monkeys had fused solidly, com- pared with only three of four mon- keys in which a collagen carrier had been utilized. Anterior Interbody Fusion Results of successful interbody fusions were reported by Hecht et al, 33 who studied the application of rhBMP-2 delivery by means of an ab- sorbable collagen sponge carrier within a freeze-dried cortical-dowel allograft in a nonhuman primate (rhesus macaque) model (Fig. 1). The rates of new-bone formation and ultimate fusion success were superior with the use of rhBMP-2 compared with autogenous cancel- lous iliac-crest graft and freeze- dried cortical-dowel allograft (Fig. 2). Successful fusion was achieved in all three animals in the rhBMP-2 group, whereas two of the three control animals had pseudarthro- ses (Fig. 3). The benefits of rhBMP-2 in stim- ulating a successful fusion response have also been demonstrated re- cently by Zdeblick et al 34 in a goat model. In that study, animals treated with cervical interbody cages filled Human Bone Morphogenetic Proteins in Spinal Fusion Journal of the American Academy of Orthopaedic Surgeons 6 with rhBMP-2 in a collagen sponge demonstrated more predictable bone growth than those treated with local bone alone. Human Studies Johnson et al 35 were among the first to use BMP in human clinical stud- ies, evaluating the role of BMPs in the treatment of femoral nonunions. Twelve patients with an average of 4.3 surgical procedures each for intractable femoral nonunions were treated with internal fixation and extracted human BMP implants. All went on to have a successful union at an average of 5 months. Johnson and Urist 36 evaluated 15 patients with posttraumatic atrophic femoral nonunions treated with a one-stage lengthening procedure (mean lengthening, 2.8 cm) involv- ing the use of an implant of allo- geneic antigen-extracted autolyzed human bone perfused with par- tially purified hBMP. Fourteen pa- tients healed primarily with no negative side effects from the allo- geneic graft material, such as infec- tion, allergic reaction, or tissue re- jection. Recently, Muschler et al 37 re- ported on the first US prospective human clinical trial of rhBMP-7, in which they evaluated its efficacy when coupled to a collagen carrier in the treatment of complicated tibial nonunions. In that study, both groups were treated with a reamed tibial nail, preparation of the nonunion site, and placement of autogenous bone graft or the rhBMP-7 device. The researchers found no clinically relevant difference in outcomes with regard to pain, return to full weight- bearing status, and avoidance of surgery between the rhBMP-7 group and the autogenous iliac-crest bone grafting group. Unfortunately, the study design did not utilize a control group with no graft material; there- fore, the effect of rhBMP-7 relative to that of the surgical procedure alone could not be discerned. An unpublished Food and Drug Administration pilot study utilizing Dan A. Zlotolow, MD, et al Vol 8, No 1, January/February 2000 7 Figure 1 Computed tomographic scans of the L5-S1 interspace in primates, obtained 3 months after attempted anterior interbody fusion with allograft. Axial (A) and sagittal (B) reconstructions of a control animal that received a cortical-dowel allograft with iliac-crest auto- graft placed inside the dowel. Minimal bone formation is demonstrated within the center of the dowel, with minimal incorporation of the cortical allograft. C and D, Images of an animal that received a cortical-dowel allograft with a collagen sponge containing rhBMP-2 placed within the dowel. Extensive bone formation is depicted within the center of the dowel, as well as extensive incorporation of the allograft with fusion at the L5-S1 interspace. (Courtesy of Jeffrey S. Fischgrund, MD, Southfield, Mich.) A B C D Figure 2 A, Histologic analysis 6 months after attempted L5-S1 fusion with cortical allo- graft and cancellous autograft within the autograft in a rhesus monkey. The allograft is still visible, and fibrous tissue is noted at the site of attempted fusion (Mallory azan, origi- nal magnification ×14). B, Similar histologic analysis of tissue from another animal that received rhBMP-2 within the cortical allograft. At the 6-month evaluation, solid fusion is demonstrated across the interspace, with normal trabecular bone. Note complete incorpo- ration of the allograft with preservation of the space available for the exiting spinal nerves. (Courtesy of Jeffrey S. Fischgrund, MD, Southfield, Mich.) A B rhBMP-2 with anterior spinal cages in the lumbar spine has reached the 1-year follow-up period. Prelimi- nary results were presented at the 1998 meeting of the North Ameri- can Spine Society and the 1999 meeting of the American Academy of Orthopaedic Surgeons. The re- searchers demonstrated 100% heal- ing by 6 months in 11 patients who received rhBMP-2 and collagen without any autograft. 6 Another human clinical trial with allograft bone dowels used as rhBMP-2 car- riers is in progress. Human clinical trials of the use of rhBMP-7 are taking place in Aus- tralia, Sweden, and Denmark, with a special focus on utilizing rhBMP-7 in lieu of autograft bone for various operative procedures in the spine. The first such study of the use of rhBMP-7 in the United States is al- ready under way. Summary Although the in vivo role of each BMP is unclear, the ability of BMPs to stimulate bone formation is no longer in question. Various studies, most recently human clin- ical trials, have demonstrated that BMPs not only can potentiate heal- ing after autologous bone grafting but also may be able to replace that procedure. It is unlikely, however, that BMPs will replace rigid fixation in spine surgery, except in cases of minimal instabil- ity. The use of BMPs in a small number of human trials suggests that rhBMPs may be safe for use in human subjects, although further investigation is necessary before widespread use can be sanctioned. Currently, the costs of BMPs (an estimated $3,000 to $5,000 per dose) limit their use to experimen- tal studies and selected cases of nonunion or those with a high probability of nonunion. However, as this technology matures, the cost will likely drop precipitously and allow BMPs to be used in other aspects of orthopaedic surgery. In theory, BMPs will be as inexpen- sive to produce as recombinant human insulin or any recombinant vaccine and will likely be as widely used. The results of the work that has been done with rhBMPs are encouraging, suggesting that the conventional techniques of spinal surgery may change dramatically in the not too distant future. Human Bone Morphogenetic Proteins in Spinal Fusion Journal of the American Academy of Orthopaedic Surgeons 8 A B Figure 3 A, Microradiograph demonstrating persistence of a cortical allograft 3 months after attempted L5-S1 fusion in a rhesus monkey. B, Microradiograph of another animal that received rhBMP-2 within the cortical allograft. Note solid fusion across the interspace, with trabecular bone formation. (Courtesy of Jeffrey S. Fischgrund, MD, Southfield, Mich.) References 1. Elima K: Osteoinductive proteins. Ann Med 1993;25:395-402. 2. Wozney JM, Rosen V, Celeste AJ, et al: Novel regulators of bone formation: Molecular clones and activities. Science 1988;242:1528-1534. 3. Einhorn TA, Lane JM, Burstein AH, Kopman CR, Vigorita VJ: The healing of segmental bone defects induced by demineralized bone matrix: A radio- graphic and biomechanical study. J Bone Joint Surg Am 1984;66:274-279. 4. Urist MR, Mikulski A, Lietze A: Solubilized and insolubilized bone morphogenetic protein. Proc Natl Acad Sci USA 1979;76:1828-1832. 5. Bauer FC, Urist MR: Human osteosar- coma-derived soluble bone morpho- genetic protein. Clin Orthop 1981;154: 291-295. 6. Boden SD, Zdeblick TA, Sandhu HS, Heim SE: The use of BMP-2 in inter- body fusion cages: Definitive evidence of osteoinduction in humans. Pre- sented at the 66th Annual Meeting of the American Academy of Ortho- paedic Surgeons, Anaheim, Calif, February 5, 1999. 7. Riley EH, Lane JM, Urist MR, Lyons KM, Lieberman JR: Bone morpho- genetic protein-2: Biology and applica- tions. Clin Orthop 1996;324:39-46. 8. Reddi AH: Bone and cartilage differ- entiation. Curr Opin Genet Dev 1994;4: 737-744. 9. Israel DI, Nove J, Kerns KM, Mout- satsos IK, Kaufman RJ: Expression and characterization of bone morphogenet- ic protein-2 in Chinese hamster ovary cells. Growth Factors 1992;7:139-150. 10. Israel DI, Nove J, Kerns KM, et al: Heterodimeric bone morphogenetic proteins show enhanced activity in vitro and in vivo. Growth Factors 1996; 13:291-300. 11. Mayer H, Scutt AM, Ankenbauer T: Subtle differences in the mitogenic effects of recombinant human bone morphogenetic proteins -2 to -7 on DNA synthesis on primary bone- forming cells and identification of BMP-2/4 receptor. Calcif Tissue Int 1996;58:249-255. 12. Honda Y, Knutsen R, Strong DD, Sampath TK, Baylink DJ, Mohan S: Osteogenic protein -1 stimulates mRNA levels of BMP-6 and decreases mRNA levels of BMP-2 and -4 in human osteosarcoma cells. Calcif Tissue Int 1997;60:297-301. 13. Chen D, Harris MA, Rossini G, et al: Bone morphogenetic protein 2 (BMP-2) enhances BMP-3, BMP-4, and bone cell differentiation marker gene expression during the induction of mineralized bone matrix formation in cultures of fetal rat calvarial osteoblasts. Calcif Tissue Int 1997;60:283-290. 14. Boden SD, Hair G, Titus L, et al: Glucocorticoid-induced differentiation of fetal rat calvarial osteoblasts is mediated by bone morphogenetic pro- tein-6. Endocrinology 1997;138:2820- 2828. 15. Boden SD, McCuaig K, Hair G, et al: Differential effects and glucocorticoid potentiation of bone morphogenetic protein action during rat osteoblast differentiation in vitro. Endocrinology 1996;137:3401-3407. 16. Takagi K, Urist MR: The role of bone marrow in bone morphogenetic pro- tein-induced repair of femoral massive diaphyseal defects. Clin Orthop 1982; 171:224-231. 17. Yasko AW, Lane JM, Fellinger EFJ, Rosen V, Wozney JM, Wang EA: The healing of segmental bone defects, induced by recombinant human bone morphogenetic protein (rhBMP-2): A radiographic, histological, and biome- chanical study in rats. J Bone Joint Surg Am 1992;74:659-670. 18. Toriumi DM, Kotler HS, Luxenberg DP, Holtrop ME, Wang EA: Mandib- ular reconstruction with a recombi- nant bone-inducing factor: Functional, histologic, and biomechanical evalua- tion. Arch Otolaryngol Head Neck Surg 1991;117:1101-1112. 19. Lovell TP, Dawson EG, Nilsson OS, Urist MR: Augmentation of spinal fusion with bone morphogenetic pro- tein in dogs. Clin Orthop 1989;243: 266-274. 20. Schimandle JH, Boden SD, Hutton WC: Experimental spinal fusion with re- combinant human bone morphogenet- ic protein-2. Spine 1995;20:1326-1337. 21. Martin GJ, Boden SD: BMP-2 reverses the inhibitory effect of ketorolac, a non-steroidal anti-inflammatory drug (NSAID), on posterolateral lumbar intertransverse process spine fusion. Presented at the 65th Annual Meeting of the American Academy of Ortho- paedic Surgeons, New Orleans, March 19, 1998. 22. David SM, Murakami T, Tabor OB, et al: Lumbar spinal fusion using recom- binant human bone morphogenetic protein (rhBMP-2): A randomized, blinded, and controlled study. Pre- sented before the International Society for the Study of the Lumbar Spine, Helsinki, Finland, June 18-22, 1995. 23. Sandhu HS, Kanim LEA, Kabo JM, et al: Evaluation of rhBMP-2 with an OPLA carrier in a canine posterolater- al (transverse process) spinal fusion model. Spine 1995;20:2669-2682. 24. Sandhu HS, Kanim LEA, Kabo JM, et al: Effective doses of recombinant human bone morphogenetic protein-2 in experimental spinal fusion. Spine 1996;21:2115-2122. 25. Boden SD, Horton WC, Martin G, Truss TR, Sandhu H: Laparoscopic anterior spinal arthrodesis with rhBMP-2 in a titanium interbody threaded cage. Presented at the 32nd Annual Meeting of the Scoliosis Research Society, St Louis, September 25-27, 1997. 26. Sheehan JP, Kallmes DF, Sheehan JM, et al: Molecular methods of enhancing lumbar spine fusion. Neurosurgery 1996;39:548-554. 27. Fischgrund JS, James SB, Chabot MC, et al: Augmentation of autograft using rhBMP-2 and different carrier media in the canine spinal fusion model. J Spinal Disord 1997;10:467-472. 28. Boden SD, Titus L, Hair G, Liu Y, Viggeswarapu M, Nanes MS: Lumbar spine fusion by local gene therapy with a cDNA encoding a novel osteoinduc- tive protein (LMP-1). Presented at the 66th Annual Meeting of the American Academy of Orthopaedic Surgeons, Anaheim, Calif, February 6, 1999. 29. Wang JC, Yoo S, Kanim LEA, et al: Gene therapy for spinal fusion: Trans- formation of marrow cells with an adenoviral vector to produce BMP-2. Presented at the 29th Annual Meeting of the Scoliosis Research Society, San Diego, Calif, September 23, 1999. 30. Sandhu HS, Kanim LEA, Toth JM, et al: Experimental spinal fusion with recombinant human bone morpho- genetic protein-2 without decortication of osseous elements. Spine 1997;22: 1171-1180. 31. Boden SD, Moskovitz PA, Morone MA, Toribitake Y: Video-assisted lat- eral intertransverse process arthrode- sis: Validation of a new minimally invasive lumbar spinal fusion tech- nique in the rabbit and nonhuman pri- mate (rhesus) models. Spine 1996;21: 2689-2697. 32. Boden SD, Moskovitz PA, Martin G: Video-assisted lateral intertransverse process arthrodesis: Validation of a new minimally invasive lumbar spinal fusion technique in the non-human primate (rhesus) model. Presented at the 12th Annual Meeting of the North American Spine Society, New York, October 25, 1997. 33. Hecht BP, Fischgrund JS, Herkowitz HN, Penman L, Toth J: The use of rhBMP-2 to promote spinal fusion in a non-human primate anterior inter- body fusion model utilizing a freeze- dried allograft cylinder. Presented at the 12th Annual Meeting of the North American Spine Society, New York, October 22-25, 1997. 34. Zdeblick TA, Ghanayem AJ, Rapoff AJ, et al: Cervical interbody fusion cages: An animal model with and without bone morphogenetic protein. Spine 1998;23:758-766. 35. Johnson EE, Urist MR, Finerman GA: Bone morphogenetic protein augmen- tation grafting of resistant femoral nonunions: A preliminary report. Clin Orthop 1988;230:257-265. 36. Johnson EE, Urist MR: One-stage lengthening of femoral nonunion aug- mented with human bone morpho- genetic protein. Clin Orthop 1998;347: 105-116. 37. Muschler GF, Perry CR, Cole JD, et al: Treatment of established tibial non- unions using human recombinant osteogenic protein-1. Presented at the 65th Annual Meeting of the American Academy of Orthopaedic Surgeons, New Orleans, March 20, 1998. Dan A. Zlotolow, MD, et al Vol 8, No 1, January/February 2000 9 . extraction schemes have been developed, but all are com- plex and do not reliably yield usable quantities of BMP, as only 1 to 2 µg of BMPs are present in a kilogram of cortical bone. Further- more,. inhibitory effects of the nonsteroidal anti-inflammatory drug ketorolac on fusion rate. 21 Dose Response Using rhBMP-2 with a collagen carrier in a canine spinal fusion model, David et al 22 showed a dose- dependent. in- cluding 58, 115, 230, 460, and 920 µg. Histologically, abundant bone formation occurred in all specimens containing rhBMP-2 by 3 months. 24 The data from this study and the one using 2,300 µg of rhBMP-2 showed