Osseointegration of autograft versus osteogenic protein–1 in posterolateral spinal arthrodesis

The Spine Journal 2 (2002) 11–24

2001 Outstanding Paper Award

Osseointegration of autograft versus osteogenic protein–1 in posterolateral spinal arthrodesis: emphasis on the comparative mechanisms of bone induction Bryan W. Cunningham, MSca,*, Norimichi Shimamoto, MDa, John C. Sefter, DOc, Anton E. Dmitriev, BSa, Carlos M. Orbegoso, MDb, Edward F. McCarthy, MDd, Ira L. Fedder, MDb, Paul C. McAfee, MDb a

Orthopaedic Research Laboratory, Union Memorial Hospital, 201 East University Parkway, Baltimore, MD 21218, USA b Department of Pathology, Union Memorial Hospital, 201 East University Parkway, Baltimore, MD 21218, USA c The Scoliosis and Spine Center, St. Joseph Hospital, 7505 Osler Drive, Baltimore, MD, USA d Department of Pathology, The Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD, USA Received 21 February 2001; accepted 10 June 2001

Abstract

Background context: Recent studies have documented increased fusion success afforded by bone morphogenetic proteins versus autogenous graft for posterolateral spinal arthrodesis. Purpose: The current study was designed to investigate the time-course maturation processes of lumbar posterolateral arthrodeses performed with Osteogenic Protein–1 (Stryker Biotech, Inc., Hopkinton, MA, USA) (rhOP-1) versus “gold standard” autograft. Study design: The primary focus of this study was to compare the histologic mechanisms of posterolateral osseointegration produced by “hot topic” growth factors. Methods: A total of 36 coonhounds were equally divided into one of four postoperative time periods of 4, 8, 12 and 24 weeks (nine animals per period). Posterolateral arthrodesis treatments included 1) autograft alone, 2) autograft plus rhOP-1, or 3) rhOP-1 alone. The treatments and animals were divided such that a value of n6 was obtained for each treatment group per time period and no one animal received the same treatment at both operative sites. Functional spinal unit (FSU) fusion status was assessed using radiographic analysis, biomechanical testing and undecalcified histopathologic and histomorphometric analyses. Results: Radiographic differences in fusion maturation between the treatment groups were evident as early as the 4-week time interval and continued through the 24-week time period. The Osteogenic Protein–1 treatments demonstrated an accelerated rate of radiographic fusion by 4 weeks, which plateaued after the 8-week time period (22% autograft, 88% autograft/rhOP-1 and 66% rhOP-1). In contradistinction, the so-called “gold standard” autograft alone treatments reached a maximum of 50% fusion by the 6-month interval. Biomechanical testing of the FSUs indicated lower flexion-extension and axial rotation range of motion levels for both rhOP-1 treatments versus autograft alone at the 8- and 12-week time periods, respectively (p.05). Histomorphometric analysis yielded no difference in the posterolateral trabecular bone area (mm2) between the three treatments (p.05), and histopathology indicated no significant histopathologic changes. The most distinctive finding in this study deals with the mechanisms of posterolateral ossification. Based on plain and polarized light microscopy, bone induction and development for the rhOP-1 treatments, with and without autograft, was the result of intramembranous ossification, whereas the process of osseointegration for autograft alone was en-

FDA Device/drug status: Approved for trauma applications in longbone fracture repair (Osteogenic Protein 1) Support in part or in whole was received from Stryker Biotech, Inc. Author B.W.C. acknowledges a financial relationship which may indirectly relate to the subject of this manuscript. * Corresponding author. Union Memorial Hospital, 201 East University Parkway, Baltimore, MD 21218. Tel.: (410) 554-2914; fax: (410) 5542408. E-mail address: [email protected] (B.W. Cunningham) 1529-9430/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S1529-9430(01)00 0 7 0 - X

12

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

dochondral bone formation. By the 24-week interval, no discernable differences in trabecular histomorphology were evident based on the different mechanisms of ossification. Conclusions: This serves as the first study to document the mechanisms of bone induction and fusion maturation between posterolateral arthrodeses treated with autograft versus rhOP-1. The histological data served to corroborate the radiographic and biomechanical findings, because the rhOP-1 treatments consistently demonstrated increased fusion rates and lower range of motion levels compared with the autograft group, particularly at the 8-week postoperative time period. The improvements in these fusion criteria for Osteogenic Protein–1 versus autograft were considered secondary to the differing mechanisms of bone induction. When implanted for posterolateral arthrodesis, rhOP-1 induces an intramembranous healing response, obviating the need for the cartilage intermediate phases found in endochondral bone development. The mechanism of increased speed and incidence of fusion using growth factors (rhOP-1) is delineated by this comprehensive study of preferential intramembranous ossification. © 2002 Elsevier Science Inc. All rights reserved. Keywords:

Bone morphogenetic protein; Posterolateral osseointegration; Animal model; Biomechanics; Histology

Introduction Since Urist’s landmark studies, which articulated the concept of specific bone morphogenetic proteins (BMPs) [1], the search for natural and synthetic bone graft substitutes continues to invigorate research in areas of bone regeneration and repair. To date, at least seven structurally related BMPs, which are members of the transforming-growth-factor  superfamily, have been isolated, purified and characterized using recombinant DNA techniques [2–6]. These osteogenic BMP factors are involved in the regulation, growth and differentiation of cartilage and bone in embryos [7–9]. Moreover, in combination with a number of cytokines and matrix components, these factors have been shown to induce a cascade of events resulting in the recruitment and differentiation of osteoprogenitor cells during the formation phase of bone remodeling [10–12]. Autologous graft, although the preferred substrate for posterolateral spinal arthrodesis [13,14], typically requires a separate surgical procedure, has a high rate of donor site complications [15–17] and is fraught with a pseudoarthrosis rate ranging from 5% to 34% [18,19]. Recombinantly pro-

duced human Osteogenic Protein–1 (rhOP-1), the focus of the current investigation, has proven efficacious in a number of in vivo animal studies. rhOP-1, also referred to as rhBMP-7, has been shown to induce ectopic bone formation in subcutaneous tissues [5,20,21], restore large diaphyseal defects in nonhuman primates [22], provide an effective bone graft substitute for achieving stable posterolateral and interbody fusions [23,24] and facilitate osteointegration of metal prosthesis [25]. Despite the many studies documenting the effectiveness of bone morphogenetic proteins and their potential to accelerate the fusion process, the mechanistic basis underlying the patterns of histologic ossification in posterolateral arthrodesis has not been clearly delineated. Moreover, from a bone induction/development viewpoint, it remains uncertain how this pathway compares with autogenous controls at equivalent postoperative time periods. With these issues at hand, the primary focus of this investigation was to compare the mechanisms of posterolateral osseointegration produced by treatments of Osteogenic Protein–1 (rhOP-1) Putty alone, OP-1 Putty in conjunction with autograft and “gold standard” au-

Fig. 1. Schematic view of the posterolateral surgical procedure.

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

13

Fig. 2. Osteogenic Protein–1 (rhOP-1) device. (Left) At the time of surgery, just before implantation, the liquid (sterile saline) and powder components of the rhOP-1 device were mixed with the carboxymethylcellulose putty additive to form the implantable material. (B) For the autograft/rhOP-1 treatments, autograft was added and thoroughly mixed within the rhOP-1 vial.

tograft alone using an in vivo spinal arthrodesis model. The time-course of fusion maturation among these three experimental groups was assessed surgically, radiographically, biomechanically and histopathologically. Materials and methods Animal model and surgical preparation All animal surgeries and experimental procedures commenced after protocol approval by the appropriate Institutional Animal Care and Use Committees. A total of 36 skeletally mature purpose-bred coonhounds (2 to 3 years old, 20 kilograms) were included in this study and followed for a period of 4, 8, 12 and 24 weeks postoperatively. After normal health status determination, each animal was sedated with intravenous (IV) injection of ketamine (10 mg/kg) and diazepam (0.25 mg/kg) anesthetic medications, followed by endotracheal intubation and general anesthesia using 1.5% to 2.0% isoflurane. With the animal positioned prone, the posterior lumbar region was shaved, aseptically prepared and draped in sterile fashion. Prophylactic antibiotics (cefazolin sodium 1 g, IV) and analgesics (Butorphenol, Fort Dodge Animal Health, Fort Dodge, IA, 125 mg/kg, IV) were administered pre- and postoperatively.

Table 1 Study Design - Schedule of Group Evaluations Treatment Groups

4 Weeks

8 Weeks

12 Weeks

24 Weeks

Autograft Control OP-1 Putty  Autograft OP-1 Putty Alone

1A (n6) 1B (n6) 1C (n6)

2A (n6) 2B (n6) 2C (n6)

3A (n6) 3B (n6) 3C (n6)

4A (n6) 4B (n6) 4C (n6)

Surgical technique and treatment groups A localization radiograph, obtained before surgical intervention, ensured exposure of the appropriate L3–L4 and L5–L6 vertebral levels. The posterior elements were exposed through a midline incision and the laminae, facet joints and transverse processes cleared of soft tissues using a Cobb elevator and electrocautery, when necessary. A bilateral muscle barrier extending bilaterally from the L4–L5 facet level served to prevent cross contamination of the treatment assignments. After exposure, the posterior cortices of the L3–L4 and L5–L6 transverse processes were decorticated using a rongeur until punctate bleeding was observed, without knowledge of the treatment group. Before implantation of the treatment procedures, the fu-

Fig. 3. Biomechanical testing. Anterior view of an operative lumbar functional spinal unit undergoing biomechanical testing in flexion/extension. To isolate the operative site, the proximal and distal ends of the motion segment were fixed with four four-point compression screws. The OptoTrak Plexiglass markers (Northern Digital, Inc., Waterloo, Ontario, Canada) are located on each vertebral level to quantify intersegmental motion.

14

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

sion bed was thoroughly irrigated to remove residual osseous debris. Two, single-level posterolateral arthrodeses at the L3– L4 and L5–L6 transverse levels were then performed. The treatment composites were placed into the right and left posterolateral gutters in direct apposition with each of the transverse processes, spanning the intertransverse space (Fig. 1). For the autograft group, 4 g of autologous tricortical iliac bone was equally distributed to the bilaterally decorticated arthrodesis site at the appropriate lumbar level (2 g per side). A tricortical wedge of iliac bone was harvested using an osteotome and prepared as morselized corticocancellous chips for re-implantation at the graft site. For the autograft/rhOP-1 treatments, 1 g autograft and 1 g of rhOP-1 were thoroughly mixed and implanted per side. The rhOP-1 Putty alone groupincluded implantation of 2 g of the material bilaterally (1 g per side) (Fig. 2, left and right). In total, three different treatments (autograft alone, rhOP-1 Putty/autograft (50/50 mixture) and rhOP-1 Putty alone) were randomized among 72 potential fusion sites (n24/treatment). The treatments and animals were divided such that a value of n6 was obtained for each treatment group per time period and no one animal received the same treatment at both operative sites (Table 1). After the surgical procedure, all incisions were closed in an interrupted fashion using 2.0 Vicryl (Ethicon, Inc., Cincinnati,

OH). Blood loss, operative times and intraoperative and perioperative complications were quantified. Observations of ambulatory activities and wound healing were monitored daily, and all animals received analgesics and prophylactic antibiotics for the first 10 days postoperatively. Animals were humanely sacrificed at the predetermined postoperative time interval using an overdose (150 mg/kg, IV) of concentrated pentobarbital sodium (concentration  390 mg/ml) The spinal column was then carefully dissected and frozen at 25 C in double-thickness plastic bags. Material specifications: recombinant human osteogenic protein-1 The implanted Osteogenic Protein–1 treatments consisted of three ingredients: 3.5 mg recombinant human Osteogenic Protein (rhOP-1), type 1 bovine bone collagen matrix and carboxymethylcellulose (CMC) (Stryker Biotech, Inc., Hopkinton, MA). These components were reconstituted at the time of implant using .9% sterile saline (Fig. 2, left and right). Recombinant human Osteogenic Protein is a highly purified protein preparation comprised of homodimers of mature polypeptides, the major species having apparent molecular weight range of 34 to 38 kilodaltons, as determined by SDS-polyacrylamide gel electrophoresis. The subunits are a mixed population of

Fig. 4. Four-week radiographic results. Analysis of the 4-week group indicated successful posterolateral fusions in (Top) 0% of autograft, (Middle) 38% autograft/ rhOP-1 and (Bottom) 22% of the rhOP-1 alone treatments. Representative anteroposterior plain film radiographs of the operative motion segments are shown.

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

species with intact and truncated amino-termini that are differently glycosylated, having molecular weights of 16 to 23 kilodaltons. The collagen matrix is an extensively purified bovine bone type 1 collagen extracted from bovine diaphyseal bone and refined to greater than 95% purity, with particles ranging in size between 75 and 425 microns. The putty additive consists of 230 mg of (CMC) sodium, which is the sodium salt of a polycarboxymethyl ether of cellulose. The CMC is used as a handling agent and serves to increase mixture viscosity.

15

B: Solid unilateral fusion: clearly solid transverse process fusion unilaterally with confluent trabeculated bone extending unilaterally from transverse process to transverse process. C: Partial union: evidence of bone growth between the transverse processes either unilaterally or bilaterally, but with lucency indicative of nonconfluent trabeculation. D: Nonunion: no evidence of bone growth between the transverse processes.

Radiological analysis

Biomechanical analysis

Status of the posterolateral spinal fusion was evaluated on the 4-, 8-, 12- and 24-week postsacrifice radiographs using the four-point grading scale documented by Lenke et al. [26]. The percentage of successful fusions (grades A or B) or pseudoarthrosis (grades C or D) were determined based on film review by three independent observers blinded to treatment groups (n6 per group  3 observations  18).

For biomechanical assessment, the frozen specimens were thawed at room temperature, and the surrounding soft tissue and musculature removed to obtain the operative ligamentous functional spinal units (L3–L4 and L5–L6). The superior half of the proximal vertebral body and inferior half of the distal body were secured into rectangular tubing foundations using four four-point compression screws. The FSU range of motion was quantified under axial rotation ( 4 Nm with 150 N compressive preload), flexion-extension ( 4 Nm) and lateral bending ( 4 Nm) nondestructive, unconstrained pure moment loading modes (Fig. 3). The axis of rotation was initially centered at the junction of the posterior one third and anterior two thirds of the operative level intervertebral disc during torsion

Grading scale A: Solid bilateral fusion: clearly solid transverse process fusion bilaterally with confluent trabeculated bone extending from transverse process to transverse process.

Fig. 5. Eight-week radiographic results. By the 8-week time interval, the success of radiographic fusion increased to (Top) 22% for autograft, (Middle) 88% autograft/rhOP-1 and (Bottom) 66% rhOP-1 alone. Representative anteroposterior plain film radiographs are shown.

16

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

and flexion-extension testing. Lateral bending adhered to the mid-coronal axis of the disc. Each test was repeated for four loading and unloading cycles at a rate of 20% full scale/second. All biomechanical testing was performed using a MTS 858 Bionix testing device (MTS Systems, Minneapolis, MN) configured with a 3020 OptoTrak Motion Analysis System and software (Northern Digital, Inc., Waterloo, Ontario, Canada). To prevent desiccation during testing, specimens were moistened with NaCl 0.9% sterile irrigation solution. Histopathological and histomorphometric evaluation Undecalcified tissue analysis Upon completion of the biomechanical analysis, the posterolateral fusion mass, including the transverse processes, were parasagittally sliced at the thickest location (left or right) using a Buehler Isomet diamond cutting system (Buehler Inc., Lake Bluff, IL). These posterolateral specimens were dehydrated in 80% ethanol, stained using the Villanueva Osteochrome Bone Stain (Polysciences, Inc., Warrington, PA), processed using undecalcified technique and embedded in polymethylmethacrylate. Using the EXAKT Microgrinding Device (EXAKT Technologies, Oklahoma City, OK), the embedded specimens were cut into 300- to 500- m-thick sections and ground and polished to 75 m in thickness. Microradiographs were then obtained us-

ing a Faxitron X-ray Unit (Faxitron X-Ray Corporation, Buffalo Grove, IL) and Konica Graphic Arts Film (Konica Imaging, U.S.A., Inc. Glen Cove, NY). The slide-mounted specimens were placed 12 inches from the beam and exposed for 2 minutes, using a technique of 42 kiloVoltspeak and 3 milliamperes while in direct contact with single-emulsion high-resolution graphics arts film. Using a BioQuant Image Analysis System (R&M Biometrics, Nashville, TN), the high-resolution microradiographs permitted histomorphometric quantification of posterolateral trabecular bone areas (mm2). The techniques for these analyses have been previously documented [27]. Decalcified tissue analysis The remaining hemisections were sectioned and evaluated by two independent board-certified cytopathologists, blinded to the treatment groups. The specimens were fixed in a 10% formalin solution, decalcified, paraffin processed and slide mounted. Using thin-sectioning microtomy, the paraffin-embedded posterolateral fusion sections were sectioned (3 to 5

m in thickness), and stained using standard hematoxylin and eosin. Using plain and polarized light microscopy, histopathological assessment for all tissues included, but was not limited to, comments on trabecular architecture, presence of collagen, as well as any signs of foreign body giant cell/gran-

Fig. 6. Twelve-week radiographic results. By 12 weeks, the success of autograft was (Top) 27%, and (Middle) the autograft/rhOP-1 83% and (Bottom) rhOP-1 alone 72% remained basically unchanged from the 8-week time period. Representative anteroposterior plain film radiographs are shown.

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

17

Fig. 7. Twenty-four-week radiographic results. By the final 24-week time interval, the success of fusion increased to (Top) 50% for autograft and was equivalent between the (Middle) 77% autograft/rhOP-1 and (Bottom) 77% rhOP-1 alone treatments. Representative anteroposterior plain film radiographs are shown.

ulomas inflammatory reactions, degenerative changes or autolysis. Moreover, the developmental ossification process of new bone, intramembranous or endochondral, was evaluated in all treatment groups for each postoperative time period. Data and statistical analysis For nondestructive biomechanical analysis, peak range of motion for each loading mode was calculated as the sum of motions (maximum displacement [millimeters] for axial compression or maximum rotation for torsion, flexion-extension and lateral bending [degrees]) occurring in the neutral and elastic zones at the fourth loading cycle. Histomorphometric data

Table 2 Plain Film Radiographic Analysis* Post-Operative Time Interval

Treatment Group Fusion Status Autograft

Autograft/rhOP-1

rhOP-1 Alone

4 Weeks 8 Weeks 12 Weeks 24 Weeks

0% (0/18) 22% (4/18) 27% (5/18) 50% (9/18)

38% (7/18) 88% (16/18) 83% (15/18) 77% (14/18)

22% (4/18) 66% (12/18) 72% (13/18) 72% (13/18)

*-Percentage of successful fusions (x/18) based on film review by three observers blinded to treatment group (n6 per group 3 observations  18).

represents the area (mm2) of trabecular bone formation within the intertransverse space. All data are shown as mean 1 standard deviation and were statistically compared using a oneway analysis of variance (ANOVA) and Student-NewmanKeuls post-hoc multiple comparison procedure. Statistical comparisons at p.05 were considered significant. In addition, a statistical power analysis was performed according to the following equations [28,29]: n = 2 × ( σ ⁄ δ ) × ( t a,ν + t 2 ( 1 – p,ν ) ) 2

ν = a(n – 1)

2

(1) (2)

where nthe sample size (the number of independent observations per group), the population standard deviation, the detectable difference, significance level, degree of freedom, t ,the t value corresponding to and , P  statistical power, a  the number of groups. Results Surgical procedures All 36 animals survived the surgery and postoperative time period without incidence of vascular, neurologic or infectious complications. A mean operative time of 80 ( 23) minutes and average blood loss of 180 ( 70) cm3 was re-

18

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

Fig. 8. (Top) Flexion-extension range of motion. This bar chart highlights the differences in mean flexion-extension range of motion between the operative treatment groups at all postoperative time intervals. The 8- and 12-week autograft/rhOP-1 and rhOP-1 alone groups indicated significantly lower flexionextension range-of-motion levels compared with the autograft alone treatments at 4- and 8-week periods (*, p.05). Power analysis of the data indicates significance at greater than .89. Bars indicate mean values, and error bars signify 1 standard deviation. (Bottom) Lateral bending range of motion. This bar chart demonstrates the decrease in lateral bending range of motion of the 8-, 12- and 24-week rhOP-1 treatments versus 4-week autograft and rhOP-1 alone groups (*, p.05). Power analysis of the data indicates significance at greater than .96. Bars indicate mean values, and error bars signify 1 standard deviation.

corded. Assessments of general appearance, ambulation and wound healing were performed for each animal. All animals were characterized as having a normal recovery throughout the 6-month postoperative period. Radiography Radiographic differences in fusion maturation among the treatment groups were evident at the early 4-week interval and continued through the 24-week time period. The Osteogenic Protein–1 treatments demonstrated an accelerated rate of radio-

graphic fusion by the 4-week time period, which plateaued after the 8-week time period (22% autograft, 88% autograft/rhOP-1 and 66% rhOP-1). The autograft alone treatments, in contrast, reached a maximum of 50% fusion by the 6-month interval. Plain film radiographic analysis of the 4-week group indicated successful posterolateral fusions in 0% of autograft, 38% autograft/rhOP-1 and 22% of the rhOP-1 alone treatments (Fig. 4). By the 8-week time interval, these values increased to 22% for autograft, 88% autograft/rhOP-1 and 66% rhOP-1 alone (Fig. 5). Similar trends were noted at 12 weeks with

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

autograft at 27% and autograft/rhOP-1 and rhOP-1 alone at 83% and 72%, respectively (Fig. 6). By 24 weeks, the autograft treatment improved to 50% (Fig. 7; Table 2). Biomechanical analysis Nondestructive biomechanical testing of the 4-week postoperative arthrodeses indicated no significant differences in peak range of motion among the three treatment groups under any loading modality (p.05). In axial rotation, the 12-week autograft/rhOP-1 treatment produced lower range of motion levels compared with autograft alone (p.05) at the same time period. Moreover, the 8- and 12-week autograft/rhOP-1 and rhOP-1 alone groups indicated significantly lower flexionextension range of motion levels compared with the 4- and 8-week autograft alone treatments (ANOVA: F3.03, p .003); however, no other comparisons were statistically significant (Fig. 8, top). Lateral bending highlighted the greatest differences in range of motion among the treatment groups. The autograft/rhOP-1 and rhOP-1 alone groups at 8, 12 and 24 weeks demonstrated lower range of motion levels than the 4week autograft and rhOP-1 alone groups (ANOVA: F4.18, p.001; Fig. 8, bottom). Power analyses indicated significance

19

levels as follows: axial rotation greater than .71, flexion/extension greater than .89 and lateral bending greater than .96. Importantly, using these nondestructive investigative methods, there was no evidence of fusion mass disruption in any specimens tested. Undecalcified and decalcified bone histomorphology Four-week treatments Histologic characterization of the 4-week autograft treatments indicated the presence of corticocancellous bone fragments within the intertransverse space, with a connective tissue interface primarily consisting of type II collagen bundles and cartilage. Based on plain and polarized light microscopy, there was significant evidence that the bone induction remodeling processes were following an endochondral pathway (Fig. 9, A). The autograft/rhOP-1 and rhOP-1 alone treatments at the same time period demonstrated the intertransverse trabecular architecture, in most cases, to be complete and composed of a dense network of woven, spindleshaped trabecular bone. In many areas, the trabeculae were in direct contact with collagen bundles undergoing mineraliza-

Fig. 9. Four-week histology results. Undecalcified histologic parasagittal sections of the 4-week specimens are shown. (A) The autograft treatments demonstrated the presence of corticocancellous bone fragments within the intertransverse region undergoing endochondral ossification. Conversely, the (B) autograft/ rhOP-1 and (C) rhOP-1 alone treatments indicated a dense network of woven trabecular bone, which was in direct contact with collagen bundles forming new trabecular structures. There was no evidence of endochondral bone formation in either rhOP-1 treatment. The mechanism of bone induction in all rhOP-1 specimens appeared to be the result of direct intramembranous ossification. TP  transverse process. (Osteochrome Villanueva Bone Stain.)

20

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

Fig. 10. Eight-week histology results. Undecalcified histologic parasagittal sections of the 8-week specimens are shown. (A) The autograft treatments demonstrate the presence of immature trabecular patterns and corticocancellous bone fragments within the intertransverse region undergoing endochondral ossification. By 8 weeks, the rhOP-1 fusion mass trabeculae are more organized and thicker compared with the 4-week specimens. In similarity to the 4-week interval, the (B) autograft/rhOP-1 and (C) rhOP-1 alone treatments indicated a dense network of woven trabecular bone. The mechanism of bone induction in both rhOP-1 treatments appeared to follow a direct intramembranous pathway. TP  transverse process. (Osteochrome Villanueva Bone Stain.)

tion. There was no evidence of a cartilaginous intermediate phase; bone induction in these 4-week specimens appeared to form by direct intramembranous ossification (Fig. 9, B and C). Eight-week treatments By 8 weeks, the corticocancellous autograft chips appear further osseointegrated into the arthrodesis sites and the presence of connective tissues, collagen and cartilage, remained evident throughout. The trabecular bone present was completely woven in structure and undergoing endochondral ossification (Fig. 10, A). In contrast, the autograft/rhOP-1 and rhOP-1 alone treatments indicated more organized and thicker fusion mass trabeculae compared with the 4-week specimens, with an extensive type II collagen network throughout. Most of the trabeculae present were composed of woven bone with thick osteoid seams, which was considered secondary to the active remodeling process. In similarity to the 4-week specimens, the bone development process for rhOP-1 treatments followed an intramembranous ossification pathway (Fig. 10, B and C). Twelve- and twenty-four-week treatments By 12 weeks, the iliac crest autograft treatments, in many cases, demonstrate complete intertransverse fusion bridges

and contain more lamellar than woven trabeculae. Moreover, there was only minor evidence of endochondral bone formation within the intended arthrodesis sites (Fig. 11, A). Those sites treated with one of the rhOP-1 groups demonstrated fusion mass trabeculae which were very well organized and primarily composed of lamellar bone, without evidence of the bone induction process (Fig. 11, B and C). By 24 weeks, confluent bridges of mature lamellar trabecular bone extending between and adjoining the operative level transverse processes are evident. The bone formation/remodeling process appears to be balanced, without indication of endochondral or intramembranous ossification. Moreover, no discernable differences in trabecular histomorphology were obvious at this time period based on the different mechanisms of early ossification (Fig. 12). All specimens reviewed are undergoing a normal healing process, without evidence of significant histopathological changes. Bone histomorphometry A total of 72 slide-mounted undecalcified histologic specimens were prepared from the posterolateral treatments.

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

Histomorphometric analysis demonstrated no statistical differences in trabecular bone formation (mm2) between the three treatment groups at any time period (p.05). The basic trend at all time periods indicated the bone formation for the autograft alone treatments as markedly lower compared with the autograft/rhOP-1 and rhOP-1 alone groups. By 8 weeks, all three treatments demonstrated considerable increases in bone formation compared with the 4-week time intervals; however, a corresponding decrease was evident at the 12- and 24-week time periods. A power analysis indicated significance level at greater than .89. Discussion This serves as the first study to document the mechanistic basis of bone induction and fusion maturation rates between posterolateral arthrodeses using “gold standard” autograft versus Osteogenic Protein–1. By definition, osteogenic protein (OP-1) or BMP-7 stimulates more osteoprogenitor cells than chondroprogenitor cells. There are three determinants of whether BMPs produce cartilage or bone: regulatory signals, response cells (MSCs) and extracellular matrix. Chondronectin

21

surface attachments on cells in cell culture can be altered to vary the amount of cartilage produced by BMP; mechanical factors, such as cutting furrows, and guanidine hydrochloride (GuHCL)–extracted bone matrix have been shown to accelerate cartilage formation. On cellulose acetate and atelocollagen, BMP-induced cartilage development is relatively scant [30]. This indicates that the surrounding environment, vascularity and amount of mechanical motion influence the expression of the BMP-7 growth factors. Several previous studies have documented the efficacy of bone morphogenetic proteins for promoting a successful spinal arthrodesis [31,32]. Schimandle et al. [33] reported on the use of rhBMP-2 in varying dosages and with different carriers for lumbar intertransverse process arthrodesis. It was concluded that arthrodesis sites using a high dose of rhBMP-2 and type 1 bovine collagen carrier were biomechanically superior to all other fusion techniques. Moreover, the radiographic and histologic findings of this study suggested that the use of rhBMP-2 and type 1 collagen carrier resulted in accelerated bone formation, consolidation and remodeling compared with autogenous bone graft alone. In similarity to these findings, Cook et al. [23], reporting on the use of rhOP-1 for

Fig. 11. Twelve-week histology results. Undecalcified histologic parasagittal sections of the 12-week specimens are shown. (A) The iliac crest autograft treatments, in many cases, demonstrate complete intertransverse fusion bridges at this time period, which contain more lamellar than woven trabeculae. Moreover, there is only minor evidence of endochondral bone formation within the intended arthrodesis sites. By 12 weeks, the rhOP-1 fusion mass trabeculae are very well organized and primarily composed of lamellar bone, without evidence of the bone induction process. TP  transverse process. (Osteochrome Villanueva Bone Stain.)

22

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

Fig. 12. Twenty-four-week histology results. Undecalcified histologic parasagittal sections of the 24-week specimens are shown. By 24 weeks, for almost all cases, the three treatment groups—(A) autograft alone, (B) autograft/rhOP-1, (C) rhOP-1 alone—demonstrated confluent bridges of mature lamellar trabecular bone extending between and adjoining the operative level transverse processes. The bone formation/ossification process appears to be balanced, without discernable differences in trabecular histomorphology resulting from different mechanisms of ossification. TP  transverse process. (Osteochrome Villanueva Bone Stain.)

canine posterolateral spinal fusions, concluded that rhOP-1 provides an effective bone graft substitute for achieving stable posterolateral fusions and is significantly quicker in time to fusion compared with autologous bone. An in vivo model is necessary to evaluate the bone remodeling process. Canines served as the experimental model for this study, because these animals are gentle, easy to care for and represent excellent models for spinal procedures. There is also extensive experience, as reported in the literature, using this model for in vivo spinal research [34–38]. For the current investigation, radiographic differences in fusion maturation were evident at the early 4-week time interval and continued through the 12-week time period, highlighting the success rate of the rhOP-1 treatments compared with autograft alone. Interestingly, the Osteogenic Protein–1 treatments demonstrated an accelerated rate of radiographic fusion by the 4-week time period, which plateaued after the 8-week time period. This was in contrast to the autograft alone treatments, which reached a maximum of 50% fusion by the 6-month interval. The biomechanical data served to corroborate the radiographic findings, because the rhOP-1 treatments consistently demonstrated lower range of motion levels com-

pared with the autograft group. The most noticeable radiographic and biomechanical transitions in rhOP-1 maturation occurred at the 8-week time period. An interesting observation in this study is that none of the treatments at any time period produced a 100% successful spinal arthrodesis rate. This is most likely attributable to the biomechanically challenging “worst case scenario” animal model (no stabilizing spinal instrumentation and no spinal brace immobilization). The histology presented the most distinctive findings for this investigation. Based on review by two independent boardcertified cytopathologists blinded to the treatments, the rhOP-1 material, implanted with and without autograft, demonstrated no evidence of endochondral bone formation. Previous studies have indicated that subcutaneous implantation of BMPs in nonorthotopic sites induce bone morphogenesis using an endochondral pathway [20,21]. The current study offers evidence that rhOP-1 results in a site-specific ossification process, because bone development in the 4- and 8-week specimens appeared to be the result of direct intramembranous ossification. In contrast, the ossification process for the autograft treatments at 4-, 8- and 12-week time periods was, in all cases, the result of endochondral bone formation. In en-

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

dochondral ossification, mesenchymal stem cells condense and differentiate into cartilage, which subsequently matures, undergoes hypertrophy, mineralizes and is invaded by osteoprogenitor cells [39]. These cells then differentiate into osteoblasts and form spicules of bone by the production and mineralization of an osteoid matrix [8]. In contrast to this mechanism, intramembranous bone formation involves direct differentiation of mesenchymal stem cells to osteoblasts [40,41]. The intramembranous bone development pathway clearly permits the osseointegration of bone to occur at an increased rate by significantly facilitating or, more likely, obviating the cartilaginous intermediate phases. These mechanisms of osseointegration are substantiated by the findings in the current investigation. Based on plain and polarized light microscopy, the 4- and 8-week autograft treatments indicated the presence of corticocancellous bone fragments within the intertransverse space, with a connective tissue interface primarily consisting of cartilage. The autograft/rhOP-1 and rhOP-1 alone treatments at the same time period demonstrated a dense network of woven trabecular bone in direct contact with collagen bundles undergoing mineralization within the intertransverse region. There was no evidence of a cartilaginous intermediate phase; bone induction in these 4- and 8-week rhOP-1 treatments was clearly the result of intramembranous ossification. Moreover, matu– ration of the rhOP-1 treatments, secondary to intramembranous bone development pathway, resulted in increased radiographical fusion, motion segment stability and histological mineralization compared with autogenous controls at equal postoperative time periods. These trends were most evident at the early 4- and 8-week time intervals. As a general histologic observation, the fusion mass size appeared to decrease as the postoperative time period increased. This observation was also confirmed based on the histomorphometric results. It is speculated that the acute healing response and large fusions, noticed at the 4- and 8-week periods, effectively remodel (according to Wolff’s law) to more appropriately address the biomechanical requirements of the operative levels by the 12- and 24-month time periods. The resulting 6-month fusion masses are more streamlined; however, the trabaculae within the fusion site were considerably thicker. Bone formation secondary to OP-1 was indistinguishable from autograft treatments. This serves as the first study to document the comparative fusion maturation rates between posterolateral arthrodeses treated with autograft and rhOP-1 in a time-course fashion. The mechanistic basis underlying the bone development process for Osteogenic Protein–1 follows an intramembranous pathway. This bone induction pathway permits increased fusion rates compared with autogenous controls, which use endochondral ossification and inherent cartilaginous intermediate phase. These differing mechanisms of ossification clearly influence the success of arthrodesis based on the radiography, biomechanics and histomorphology at the early time periods. It is important to note that although the pathways, endochondral or intramembranous, leading to the posterolateral fusion

23

mass incorporation are distinctly different, the trabecular architecture of the 6-month fusion sites are indistinguishable. The potential advantages of this material for posterolateral spinal arthrodesis are increased fusion success, decreased time to fusion, improvement in clinical outcome and elimination of the iliac crest surgical procedure and associated patient morbidity. Further research is necessary to delineate and specifically characterize the histology, immunocytochemistry and gene expression during osseointegration of Osteogenic Protein–1 versus autograft in posterolateral spinal arthrodesis. Acknowledgments The authors would like to thank the University of Maryland Biotechnology Institute, Animal Core Facility and Research Staff for their assistance with the animal husbandry procedures. We would also like to thank the Departments of Cytopathology at The Johns Hopkins Hospital and Union Memorial Hospital for their independent analysis. References [1] Urist MR. Bone: formation by autoinduction. Science 1965;150:893–9. [2] Boden SD. Bioactive factors for bone tissue engineering. Clin Orthop 1999;367S:S84–94. [3] Luyten FP, Cunningham NS, Ma S, et al. Purification and partial amino acid sequence of osteogenin, a protein initiating bone differentiation. J Biol Chem 1989;264:13370–80. [4] Ozkaynak E, Rueger DC, Drier EA, et al. OP-1 cDNA encodes an osteogenic protein in the TGF-beta family. European Molecular Biology Association (EMBA) J 1990;9:2085–93. [5] Sampeth TK, Coughlin JE, Whetstone RM, et al. Bovine osteogenic protein is composed of dimers of OP-1 and BMP-2A, two members of the transforming growth factor-beta superfamily. J Biol Chem 1990;265:13198–205. [6] Sampeth TK, Reddi AH. Dissociative extraction and reconstitution of extracellular matrix components involved in local bone differentiation. Proc Natl Acad Sci USA 1981;78:7599–603. [7] Celeste AJ, Iannazzi JA, Taylor RC, et al. Identification of transforming growth factor  family members present in bone-inductive protein purified from bovine bone. Proc Natl Acad Sci USA 1990;87: 9843–7. [8] Reddi AH. Cell biology and biochemistry of endochondral bone development. Collagen Relat Res 1981;1:209–26. [9] Wozney JM, Rosen V, Celeste AJ, et al. Novel regulators of bone formation: molecular clones and activities. Science 1988;242:1528–34. [10] Cox K, McQuaid DP, Rosen V. Use of cell capture chambers to study mesenchymal stem cells. Responses to recombinant human bone morphogenetic protein-2. J Bone Miner Res 1993;8(suppl 1):S179. [11] Katagiri T, Yamaguchi A, Ikeda T, et al. The non-osteogenic mouse pluripotent cell line, C3H10T1/2, is induced to differentiate into osteoblastic cells by recombinant human bone morphogenetic protein-2. Biochem Biophys Res Commun 1990;172:295–9. [12] Takuwa Y, Ohse C, Wang EA, et al. Bone morphogenetic protein-2 stimulates alkaline phosphatase activity and collagen synthesis in cultured osteoblastic cells, BC3T3-E1. Biochem Biophys Res Commun 1991;174:96–101. [13] Heiple KG, Chase SW, Herndon CH. A comparative study of the healing process following different types of bone transplantation. J Bone Joint Surg 1963;45A:1592. [14] Wilson PD, Lance EM. Surgical reconstruction of the skeleton following segmental resection of bone tumors. J Bone Joint Surg 1965;47A:1629.

24

B.W. Cunningham et al. / The Spine Journal 2 (2002) 11–24

[15] Cockin J. Autologous bone grafting. Complications at the donor site. J Bone Joint Surg [Br] 1971;53:153. [16] Damien CJ, Parsons JR. Bone graft and bone graft substitutes. A review of current technology and applications. J Applied Biomaterials 1991.26;2:187–208. [17] Younger EM, Chapman MW. Morbidity at bone graft donor site. J Orthop Trauma 1989;3:192–5. [18] Eie N, Solgaard T, Kleppe H. The knee-elbow position in lumbar disc surgery: a review of complications. Spine 1983;8:897–900. [19] Zdeblick TA. A prospective, randomized study of lumbar fusion. Preliminary results. Spine 1993;18:983–91. [20] Reddi AH, Cunningham NS. Initiation and promotion of bone differentiation by bone morphogenetic proteins. J Bone Min Res 1993;8: S2:2499–502. [21] Ripamonti U, Crooks J, Matsaba, T, Tasker J. Induction of endochondral bone formation by recombinant human transforming growth factor-beta2 in the baboon (Papio ursinus). Growth Factors 2000;17(4):269–85. [22] Cook SD, Baffes GC, Wolfe MW, et al. Recombinant human osteogenic protein-1 (rhOP-1) heals segmental long bone defects in nonhuman primates. Trans Orthop Res Soc 1993;18:484. [23] Cook SD, Dalton JE, Tan EH, Whitecloud TS, Rueger DC. In-vivo evaluation of recombinant human osteogenic protein (rhOP-1) implants as a bone graft substitute for spinal fusions. Spine 1994;19:1655–63. [24] Cunningham BW, Kanayama M, Parker LM, et al. Osteogenic protein versus autologous interbody arthrodesis in the sheep thoracic spine. Spine 1999;24:509–18. [25] Rutherford RB, Sampath TK, Rueger DC, et al. Use of bovine osteogenic protein to promote rapid osseointegration of endosseous dental implants. Int J Oral Maxillofac Implants 1992;7:297–301. [26] Lenke LG, Bridwell KH, Bullis D, Betz RR, Baldus C, Schoenecker P. Results of in situ fusion for isthmic spondylolisthesis. J Spinal Disord 1992;5:433–42. [27] Cunningham BW, Oda I, Haggerty CJ, et al. The use of bioactive glass for posterolateral spinal arthrodesis and iliac crest donor site repair—an in vivo sheep model. Trans Orthop Res Soc 1999;24:537. [28] Glantz SA, Slinker BK. Primer of applied regression and analysis of variance. New York: McGraw-Hill, 1990:710–20.

[29] White AA, Panjabi MM. Clinical biomechanics of the spine, 2nd ed. Philadelphia: JB Lippincott, 1990;1–83,635–96. [30] Katoh R, Urist MR. Surface adhesion and attachment factors in bone morphogenetic protein-induced chondrogenesis in vitro. Clin Orthop Rel Res 1993;295:295–304. [31] Akihito M, Tamaki T, Kawakami M, Hashizume H, Yoshida M, Sakata R. Experimental spinal fusion using sintered bovine bone coated with type I collagen and recombinant human bone morphogenetic protein–2. Spine 1999;24:1863. [32] David SM, Gruber HE, Meyer RA, et al. Lumbar spinal fusion using recombinant human bone morphogenetic protein in the canine. A comparison of three dosages and two carriers. Spine 1999;24(19): 1973–9. [33] Schimandle JH, Boden SD, Hutton WC. Experimental spinal fusion with recombinant human bone morphogenetic protein-2 (rhBMP-2). Spine 1995;20:1326–37. [34] McAfee PC, Farey ID, Sutterlin CE, Gurr KR, Warden KE, Cunningham BW. Device related osteoporosis with spinal instrumentation. A canine model. Spine 1989;14(9):919–26. [35] McAfee PC, Farey ID, Sutterlin CE, Gurr KR, Warden KE, Cunningham BW. The effect of spinal implant rigidity on vertebral bone density: a canine model. Spine 1991;16(6S):S190–7. [36] Sandhu HS, Kanim LEA, Kabo JM, et al. Effective doses of recombinant human bone morphogenetic protein-2 in experimental spinal fusion. Spine 1996;21(18):2115–22. [37] Sandhu HS, Kanim LEA, Kabo JM, et al. Evaluation of rhBMP-2 with an OPLA carrier in a canine posterolateral (transverse process) spinal fusion model. Spine 1995;20(24):2669–82. [38] Sandhu HS, Kanim LEA, Toth JM, et al. Experimental spinal fusion with recombinant human bone morphogenetic protein-2 without decortication of osseous elements. Spine 1997;22(11):1171–80. [39] Owen M. The origin of bone cells in the postnatal organism. Arthritis Rheum 1980;23:1073–80. [40] Bernard GW, Pease DC. An electron microscopic study of initial intramembranous osteogenesis. Am J Anat 1969;125:271–90. [41] Marvaso V, Bernard GW. Initial intramembraneous osteogenesis in vitro. Am J Anat 1977;149:453–68.