- Email: [email protected]
Long-term effects of local growth factor (IGF-I and TGF-β1) treatment on fracture healing
Journal of Orthopaedic Research ELSEVIER
Journal of' Orthopaedic Research 22 (2004) 514-519 www.elsevier.com/locate/orthres
Long-term effects of local growth factor (IGF-I and TGF-p l ) treatment on fracture healing * A safety study for using growth factors Gerhard Schmidmaier *, Britt Wildemann, Daniel Ostapowicz, Frank Kandziora, Richard Stange, Norbert P. Haas, Michael Raschke Dipirtiiiriif
of' Trrrzollu
ciid
Riwmstnrcriw Sitrgi,rj,, Hurllholr/t Uiiiwrsitj, ( J / Biviin. Charif&,Curllpus Virclioit~C h i c , Augu.strrrh~rrgcrPlirr: I, Bcrliir D- 133.53, Gerillurly
Accepted 26 September 2003
Abstract
Previous studies showed that growth factors dramatically stimulate healing processes in bone. However, the long-term effect of locally applied growth factors on fracture healing remains unclear. In considering the safety of using growth factors, it is necessary to elucidate that after initial stimulation, the effect stops and the result is a normally healed tissue. Therefore, the purpose of the present study was to investigate the long-term time course of healing processes during growth factor (GF) stimulated and unstimulated fracture healing in a closed tibia1 fracture model in rats. A well established local drug delivery system was used. IGF-I (50 pg) and TGF-PI (10 pg) were locally applied using a 10 pin thin polylactide (PDLLA) coating on intramedullary implants. The biomechanical and histomorphometrical results demonstrated a significant stimulation of the fracture healing due to the locally applied growth fxtors compared to control at days 28 and 42 in agreement with the literature. At the last time point, 84 days after fracture, no differences were measurable in the biomechanical tcsting and the callus composition bctween the groups. The callus was consistently in the late phase of remodeling with no remaining cartilage. In conclusion, local growth factor application enhances the healing in the early phase without alteration of the normal healing process. 0 2004 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. Ki,jiwrt/.s:
Long-term effect; Drug delivery; IGF-I: TGF-PI: Fracture healing
Introduction In orthopaedic surgery, great interest continues in efforts to shorten fracture healing time, to increase the rates of unions o r complete regeneration, a n d to treat non-unions successfully. Since the discovery of Marshal Urist in 1965 that demineralized bone can induce ectopic bone formation [30] and that growth factors are responsible for this induction, the research o n this topic has increased dramatically [19]. The number of identified growth factors involved in bone healing processes increased in recent decades. These factors have been characterized [ 151, and the distribution a n d quantifica-
Parts of this study have been presented at the ORS 2003. *Corresponding author. Tel.: +49-30-450-552043: fax: +49-30.45@ 552943. E-riluil ur/c/ws.s: [email protected] (G. Schmidmaier).
tion in the callus tissue [5,8,10,26,34], the mode of action on different cell types [7,11,16,18,29], the signaling cascades [14,12,9] have been analyzed. F o r growth factors such as insulin-like growth factor-I (IGF-I) and transforming growth factor-beta1 (TGF-Pl), a stimulating effect on bone healing was clearly demonstrated in several studies using different animal models [2,4,23,28]. A benefit in using these growth factors is that they d o not induce ectopic bone formation, which has been described for BMPs [35]. Local application of growth factors to enhance bone healing has been accomplished through various carrier systems o r by direct injection. Injection o r local release of TGF-P from tricalcium-phosphate o r hydroxyapatite stimulated bone healing [ 131. Continuous infusion of IGF-I into the arterial supply of the right hind limb of rats affected cortical a n d trabecular bone formation [23]. A critical-size calvarial defect in rats showed improved healing when treated with I G F using a n osmotic
0736-02666 - see front matter 0 2004 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved doi:l0. 1016/j.orlhres.2003,09.0~9
infusion pump [28]. BMP-2 delivered from collagen sponges enhanced fracture healing in a goat tibial fracture model [32] and in an ulnar osteotomy model in the rabbit [6]. These application techniques, however, require the opening of the fracture, direct injections, or the placement of further devices and may have possible side effects due to the use of carriers like bovine collagen ~251. Despite the very promising results of locally applied growth factors on bone healing, concerns due to the long time effects of the factors remains. For example, a misbalance of TGF-PI can have undesirable effects; high local doses can be associated with unresolved inflammation [31]. Studies investigating the side effects of systemically delivered growth factors revealed severe side effects [27,33]. It is unclear if the processes stimulated by locally applied growth factors follow a physiological pattern. A permanent long-term stimulation of bone healing could result in an abnormal bone structure. The normal remodeling of the fracture callus is an important safety issue for clinical application. The hypothesis of the present study was that locally applied growth factors enhance fracture healing in the early phase, but eventually lead to a normal remodeling of the callus resulting in a comparable bony srructure compared to the unstimulated fractures. To investigate this hypothesis, the effects of IGF-I and TGF-PI on different phases of fracture healing were examined. A previous study showed growth factor stimulation resulted in an increased torsional stability compared to the control group and the unfractured contralateral tibia at 28 and 42 days after fracture [20]. However, no data are available concerning longer term effects. In the present study, the progress of rat fracture healing was traced radiologically, and biomechanical torsional testing and histomorphometric analyses were performed after 28, 42 and 84 days.
Materials and methods Five-month-old female Sprague Dawley rats (Harlan-Winkelmann, Borchen, Germany) ( n = 191) were used for the study. After sedation of the animals with Isofluran (Forene') and intraperitoneal anesthesia with a mixture of ketaminhydrochloride (100 mglml) (80 mg/kg body weight) and xylazin 2'%, (12 mg/kg body weight). a closed midshaft fracture of the right tibia was produced using a special fracture device described previously [20]. After closed reposition, the tibiae were intramedullary stabilized with coated versus uncoated titanium Kirschner-wires (1.0 mm diameter, 3.0 cm length). The growth factors recombinant human IGF-I (R&D System, Wiesbaden. Germany) (5% w/w) and recombinant human TGF-Dl (R&D Systems. Wiesbaden, Germany) (1% w/w) were locally applied using poly(D.L-lactide) (Boehringer. Ingelheim, Germany) as a drug carrier coated on the wires. Three groups were investigated: Group 1: implants uncoated. Group 11: implants coated with PDLLA alone. Group 111: implants coated with IGF-I (50 pg) and TGF-Dl (10 pg).
The properties of the PDLLA coating and the release characteristics of the incorporated growth factors were described previously [22]. Briefly, approximately 50% of the incorporated growth factors were released within the first 48 h followed by a sustained release of a further 30% over the next 40 days. The 10 pm thin coating displayed a high degree of mechanical stability on the metallic implant. After healing periods of 28, 42, and 84 days, 20 rats of each group were killed. and the tibiae were prepared for biomechanical testing or histology. The Animal Experimentation Ethics Committee approved all experiments. Throughout the 84 day observation period, radiographs were taken in posterior-anterior and lateral views using Microvision "C Mammography films (Sterling Diagnostic, Newark, D E ) and a Mobilett Plus X-ray unit (Siemens AG, Erlangen. Germany). The bridging of the fracture callus at days 28.42, and 84 was described using following parameters by two independent observers: (a) complete bridging (four cortices were bridged): ( b ) unilateral bridging (1-3 cortices were bridged); (c) n o bridging (no cortex was bridged). For mechanical testing. ten animals were randomly chosen from those killed at each group/time point. and both tibiae were dissected free from soft tissue. The unfractured side served as a control. and data were expressed in percent to the unfractured tibia. After careful dissection of the bones and removal ofthe K-wire, the proximal and distal ends were embedded into two molds with bone cement (Beracryl. Troller, Fullenbach. Switzerland). Each mold was connected to a pivoted axis. Torsion was applied by a material testing machine (Zwick 1455. Ulm, Germany) loading one of the pivoted axes by a lever. The bone was preloaded by an axial force of 5 N, and constant linear displacement (2 mm/min) was applied by the testing machine. The free axis was connected with a load cell (F,,,, = 50 N. HBM. Germany) that determined the torsional force. For histological and histomorphometrical analyses, the other 10 fractured tibiae from each group/time point were fixed for 2 days in 10'%,normal buffered formaldehyde, followed by dehydration in ascending concentrations of ethanol. and embedded undecalcified in methylmethacrylate (Technovit 9 100, Heraeus, Wehrheim, Germany). Longitudinal sections in a sagittal plane were cut at 5 pm with a Leica SM 2500s microtome (Bensheim, Germany) with a 40" stainless steel knife. Von Kossa and a combination of safranin-orange and light-green stains were used. Histomorphometric parameters in the fracture callus were measured using a Leica DM-RB (Bensheim. Germany) microscope and an image analysis system (KS 400. Zeiss, Gottingen. Germany). Parameters were measured at a magnification of 1.6 times, and the following structural indices were calculated: tibial diameter at the fracture gap (Ti' Dm), including Ct . Wi and Ma Dm (mm), bone density of the corticalis. defined as mineralized area/cortical bone area: Md . Ar/Ct B . Ar (I%>): callus area, CI . Ar (mm'); mineralized area of the CI . Ar, Md Ar/CI. Ar (I%,): and cartilage area of the C1.Ar. Cg.ArlC1.Ar ('%I), where Ar=area. B = bone, Cg = cartilage, Ct = corticalis. Dm =diameter, Ma = marrow, Md = Mineralized, and Wi = Width, all based on Parfitt et al. [ 171. and CI = callus. The region of interest (ROI) was defined by the individual tibial diameter (baseline). The callus was divided into the proximal and distal part, and 1.5 baseline lengths were used to define the ROI of proximal and distal callus halves. The total diameter of the callus was included in the ROI. Histomorphometric comparisons between the groups were based on the mineralized volume of the cortices, the area of the whole calluses, and the mineralized and cartilaginous volume of the calluses
PI]. StntlTticciI unulj,sis
The animals were randomized in a blinded manner for radiographic evaluation and for the histological and biomechanical investigations. Data were compared using one way ANOVA for independent samples. The radiological score was analyzed using the %'-test. Both tests were controlled with Bonferroni correction. Interobserver variability for the radiographic evaluation was determined using kappa statistics. Statistical differences were defined at a 95% confidence level. The values are given as mean k standard deviation. SPSS (release 10.0; SPSS Chicago, Illinois) software supported the statistical evaluation.
Results
Eleven animals were excluded from the study due to death during anesthesia or a complex tibia1 fracture at day 0. Differences in the consolidation were seen at 28 and 42 days after fracture, but not at 84 days. The uncoated group I showed a reduced consolidation compared to the PDLLA group I1 and the growth factor group I11 after 28 and 42 days (Fig. 1, Table 1). All groups showed complete consolidation by 84 days. Mechanical testing of tibiae 28 days after fracture caused failure mainly in the callus region in the uncoated and PDLLA-coated groups or adjacent to the fracture callus in the growth factor group. At 42 days, the fractured tibiae of the uncoated group failed mainly in the callus, while those of the PDLLA and growth factor groups failed in the distal or proximal part of the tibia. All three groups failed outside the callus region 84 days after fracture.
The fractures of the uncoated group revealed a significantly ( p < 0.05) lower maximum load and stiffness 28 and 42 days after fracture compared to the other two groups. The growth factor group I11 showed a significantly higher stiffness and load at 28 and 42 days than the PDLLA-coated group. At 84 days, no significant differences were measurable among any of the three groups. The maximum load and the torsional stiffness were approximately 50's) larger than the unfractured tibia (Fig. 2). At 28 and 42 days, the calluses of all three groups were composed of fibroblasts, cartilage cells, and newly formed trabecular bone (Fig. 3). At 28 days, the healing process was in the phase of enchondral ossification in all three groups. By 84 days, the calluses were in the phase of remodeling, and no cartilage was seen in any animal. The histomorphonietry supported the histological evaluation. Control parameters such as the diameter of the tibiae (baseline) and the mineralized area of the cortices showed no significant differences between the
Fig. I . Radiographs (lateral) of right tibiae 28. 42, and 84 days after fracture and intramcdullary stabiliLation with coated versus uncoated titanium K-wires. The progressed consolidation of the fracture i n the growth f x t o r group (IGF-I + TGF-al) is clearly recognizable compared t o the control group. At 84 days after fracture, the fractures of the control and the gt-owtli factor rats were completely consolidated.
Table I Consolidation of the fractures
Complete bridging Incomplete bridging No bridging * p < 0.05
(L2).
IGF-I + TGF-PI
PDLLA
Uncoated 28
42
84
28
42
84
28
42
84
I
3
10
3
5
I0
5'
9.
10
2
-7
0
3
3
0
4
1
0
7
5
0
4
-7
0
I.
0'
0
C. Sclinriiltnuirr et ul. / Journal of Ortlropurdic Reseurch 22 (2004) 514- 51 9
517
groups at any of the time points (Table 2). At the earliest time point, the growth factor treated animals had significantly less cartilage in the callus compared to group I and I1 (Fig. 4a). This difference was also seen 42 days after fracture. Additionally, the callus size was reduced and the mineralized area of the callus was significantly enhanced in the growth factor group I11 compared to group I and I1 (Fig. 4b). At this time point, the PDLLA group also had significantly less cartilage compared to group I. No differences in the callus composition were measurable 84 days after fracture (Table 2). The callus sizes in all three groups were comparable and slightly reduced compared to those found 42 days after fracture. The mineralization was increased in all three groups compared to the 42 day results, and no cartilage was measurable.
Discussion
Fig. 2. Results of torsion testing 28, 42, and 84 days after fracture in comparison to the contralateral unfractured tibiae (in %). At 28 and 42 days, the largest loads (a) and torsional stimnesses (b) were measured in ) by the PDLLA the growth factor group ( I G F - I + T G F - ~ Ifollowed group. No significant differences were found 84 days after fracture.
An enhancement of the healing process due to the application of IGF-I (50 pg) and TGF-PI (10 pg) growth factors delivered locally from a PDLLA-coated implant [22] was measurable 28 and 42 days after fracture as shown before [ ~ o I~. ~ d showed i enhanced ~ ~ con-~ solidation of the fractures after growth factor treatment, and the torsional stiffness and maximum load increased significantly in the growth factor group compared to
Fig. 3. Histological sagittal sections of tibiae 28, 42, and 84 days after fracture stained with Safranin O/light green. The mineralized proportion stained in green, the cartilage in red. A progressive callus remodeling is visible between the growth factor group and the control group.
Table 2 Histomortlhometric results 84 days after fracture
Tibia diameter (base) (mm) Mineralized Adcortex Ar (%I) Periosteal callus Adbase (mm) Mineralized Ar/Ps. CI .Ar (%I) Cartilage Ar/Ps. CI .Ar ('%)
No significant differences between the groups.
Uncoated
PDLLA
IGF-I + TGF-el
2.7 f 0.3 97.2 f 1 .O 3.8f1.7 91.2 f 5.9 0
2.6 ? 0.1 97.52 1.1 2.7 ? 1.O 96.2 2 3.1 0
2.7 f 0.3 97.4 f 0.7 3.7f 1.1 94.0 f 5.1 0
~
~
518
G. Schmidniuicr el uf. I Journul of' Ortltopurriic Reseurch 22 (2004) 514-51 9
Fig. 4. Histomorphometric analyses of the callus composition 28, 42, and 84 days after fracture: (a) the cartilage area in percent t o the total callus area and (b) the mineralized area in percent to the total callus area (where Ar = area, CI =callus. Dm = diameter, Ps = periosteal, and Ti =tibia. * Denotesp < 0.05 between groups at the same time point).
those of the control and the PDLLA-group and of the unfractured contralateral side. The histomorphological and histomorphometrical analyses supported the biomechanical results showing a significantly accelerated callus remodeling in the growth factor group. However, at the longest time point, 84 days after fracture, no differences were found in any of these measures among the three groups. All investigated fractures were completely consolidated. The strength of the fractured bone was consistently higher by approximately SO'%, compared to the unfractured bone. Callus was detectable that was composed only of mineralized tissue without any cartilage. No differences in callus size and mineralization were observed among the groups. This lack of a long-term effect is important for the safety of using growth factors to stimulate bone healing. A concern in using growth factors is that they might result in an enhanced tissue formation without normal remodeling. This study clearly demonstrates that the growth factors affect the healing in the earlier time points, but that at the late time point the fracture is remodeled comparable to the control groups. Other studies investigated the effect of locally applied growth factors in the early and middle phase of healing [3,4,13,32,36]. All these studies showed a stimulating effect of the applied growth factors on the healing. This study reaffirms that fracture healing is accelerated due to the exogenously applied growth factors without alteration of the physiological processes [20]. At the longest
time point, the callus was in the late phase of remodeling. The cartilage observed at the earlier time points was remodeled, and the callus size was significantly reduced compared to the earlier time points. Furthermore, no side effects such as neoplasia or deficient bone production were found. Using a critical size defect, Baltzer et al. were able to demonstrate better healing due to adenoviral transfer of BMP-2 [l]. However, 12 months after surgery and BMP2 application, the biomechanical stability was still increased compared to a control group. Their main focus was on the effect of the growth factors. Several investigations have been focused on the distribution of different factors and receptors during the very early phases of healing [24,26,34]. Our results are in accordance with a study investigating bone closure of scull defects [2]. The treatment of non-healing scull defects with TGF-P1 lead to a stimulation of bone matrix deposition at the earlier investigated time points without alteration of the normal remodeling at the late healing phase. In summary, the investigation of local growth factor application at different time points with different methods revealed an accelerated healing at the early phases (enchondral ossification and the beginning of remodeling). However, at the last time point investigated, the phase of remodeling, no significant differences occurred in torsional stiffness and callus composition with growth factor treatment. The radiological score showed a complete consolidation of the fractures in all investigated animals. These results lead to the conclusion that the local application of growth factors accelerates the healing in the early phase without alteration of the physiological processes and results in a normal remodeling of the callus.
Acknowledgements This study was supported by Deutsche Forschungsgemeinschaft DFG (Schm 1436/1-1).
References [I] Baltzer AW, Lattermann C, Whalen JD. Wooley P, Weiss K. Grimm M, et al. Genetic enhancement of fracture repair: healing of an experimental segmental defect by adenoviral transfer of the BMP-2 gene. Gene Ther 2000;7:73&9. [2] Beck L, Amento E, Xu Y , Deguzman L, Lee W, Nguyen T, et al. TGF-beta1 induces bone closure of skull defects-temporal dynamics of bone formation in defects exposed to rhTGF-betal. J Bone Miner Res 1993;8:75341. [3] Blokhuis TJ, den Boer FC, Bramer JA, Jenner JM, Bakker FC, Patka P, et al. Biomechanical and histological aspects of fracture healing, stimulated with osteogenic protein-I. Biomaterials 2001; 22~725-30.
[4] Blunienfeld I, Srouji S. Lanir Y, Laufer D. Livne E. Enhancement of bone defect healing in old rats by TGF-beta and IGF-I. Exp Gerontol 2002;37:553-65. [S] Bourque WT, Gross M, Hall B K . Expression of four growth factors during fracture repair. Int J Dev Biol 1993;37:573-9. [h] Bouxsein ML. Turek TJ, Blake CA. D'Augusta D. Li X . Stevens M. et al. Recombinant human bone morphogenetic protein-2 accelerates healing in a rabbit ulnar osteotomy model. J Bone Joint Surg Am 2001;83:1219-30. [7] Canalis E. Effect of insulinlike growth factor I on DNA and protein synthesis in cultured rat calvaria. J Clin Invest 1980;66: 709 19. [8] Cho T.J. Gerstenfeld LC, Einhorn TA. Differential temporal expression of members of the transforming growth factor beta superfamily during murine fracture healing. J Bone Miner Res 2002;17:513-20. [9] Claeys I, Siinonet G, Poels J, Van Loy T, Vercammen L, De Loof A. et al. Insulin-related peptides and their conserved signal transduction pathway. Peptides 2002;23:807-16. [lo] Eingartner C , Coerper S, Fritz J, Gaissmaier C , Koveker G, Weise K . Growth factors in distraction osteogenesis. Immuno-histological pattern of TGF-beta1 and IGF-I in human callus induced by distraction osteogenesis. Int Orthop 1999;23:253-9. [ I I ] Hock J. Centrella M, Canalis E. Insulin like growth factor I has independent effects on bone matrix formation and cell replication. Endocrinology 1988:122:25UiO. [I21 Jones JI, Clemmons DR. Insulin-like growth factors and their binding proteins: biological actions. Endocr Rev 1995;16:3-34. [I31 Lind M. Growth factor stimulation of bone healing. Effects on osteoblasts, osteoinies, and implants fixation. Acta Orthop Scand Suppl 1998;283:2-37. [I41 Massague J. TGF-beta signal transduction. AIIIIURev Biochem 1998;67:753-91. [I 51 Massague J. The transforming growth factor-beta family. Annu Rev Cell Biol 1990;6:597-641. (161 Matsumura T. Whelan MC, Li XQ. Trippel SB. Regulation by IGF-I and TGF-beta1 of swarm-rat chondrosarcoma chondrocytes. J Orthop Res 2000:18:3S1-5. [17] Parfitt AM, Drezner MK. Glorieux FH, Kanis JA, Malluche H, Meunier PJ. et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphonietry Nomenclature Committee. J Bone Miner Res 1987; 2:595-610. [IS] Pfeilschifter J. Oechsner M. Naumann A, Gronwald R, Minne H, Ziegler R. Stimulation of bone matrix apposition in vitro by local growth factors: a comparison between Insulin-like growth factor I, platelet-derived growth factor and transforming growth factor beta. Endocrinology 1990;127:69-75. [I91 Schliephake H. Bone growth factors in maxillofacial skeletal reconstruction. Int J Oral Maxillofac Surg 2002;31:469-84. [20] Schmidmaier G, Wildemann B, Bail H, Lucke M, Fuchs T, Stemberger A, et al. Local application of growth factors (insulinlike growth factor-l and transforming growth factor-l31) from a biodegradable poly(o,l.-lactide) coating of osteosynthetic implants accelerates fracture healing in rats. Bone 2001;28:341-50.
[21] Schmidmaier G, Wildeinann B. Cromme F, Kandziora F, Haas NP, Raschke M. BMP-2 coating of titanium implants increases biomechanicdl strength and accelerates bone remodeling in fracture treatment. Bone 2002;6:618-22. [22] Schmidmaier G, Wildemann B, Stemberger A, Haas NP, Raschke M . Biodegradable poly( D , L-lactide) coating of implants for continuous release of growth factors. J Biomed Mater Res Appl Biomat 2001;58:449- 55. [23] Spencer EM, Liu CC, Si EC, Howard GA. In vivo actions of insulin-like growth factor-I (IGF-I) on bone formation and resorption in rats. Bone 1991:12:21 6. [24] Steinbrech DS, Mehrara BJ, Rowe NM, Dudziak ME, Luchs JS. Saadeh PB, et al. Gene expression of TGF-beta, TGF-beta receptor, and extracellular matrix proteins during membranous bone healing in rats. Plast Reconstr Surg 2000;105:2028-38. [25] Takaoka K, Koezuka M, Nakahara H. Telopeptide-depleted bovine skin collagen as a carrier for bone morphogenetic protein. J Orthop Res 1991:9:902 7. [26] Tatsuyama K, Maezawa Y, Baba H, lmamura Y, Fukuda M. Expression of various growth factors for cell proliferation and cytodifferentiation during fracture repair of bone. Eur J Histochem 2000;44:269-78. 1271 Terrell TG, Working PK, Chow CP, Green JD. Pathology of recornbinant human transforming growth factor-beta 1 in rats and rabbits. Int Rev Exp Pathol 1993;34B:43-67. (281 Thaller SR, Dart A, Tesluk H. The effects of insulin-like growth factor- 1 on critical-size calvarial defects in Sprague-Dawley rats, Ann Plast Surg 1993:31:429~33. [29] Tsukazaki T, Usa T, Matsumoto T, Enomoto H, Ohtsuru A. Namba H, et al. Effect of transforming growth factor-beta on the insulin-like growth factor-] autocrinelparacrine axis in cultured rat articular chondrocytes. Exp Cell Res 1994215% 16. [30] Urist M R . Bone formation by autoinduction. Science 1965;ISO: 893-9. [31] Wahl S M . Transforming growth factor beta: the good, the bad, and the ugly. J Exp Med 1994:180:1587-90. [32] Welch R D , Jones AL, Bucholz RW, Reinert CM. Tjia JS, Pierce WA, et al. Effect of recombinant human bone morphogenetic protein-2 on fracture healing in a goat tibia1 fracture model. J Bone Miner Res 1998;13:1483-90. [33] Wilton P. Treatment with recombinant human insulin-like growth factor I of children with growth hormone receptor deficiency (Laron syndrome). Acta Paediatr Suppl 1992:383: 137-42. [34] Yu Y, Yang JL, Chapman-Sheath PJ, Walsh WR. TGF-beta, BMPS, and their signal transducing mediators, Smads, in rat fracture healing. J Biomed Mater Res 2002:60:392 7. [35] Yoshida K, Bessho K, Fujimura K, Kusumoto K. Ogawa Y, Tani Y, et al. Osteoinduction capability of recombinant human bone morphogenetic protein-2 in intramuscular and subcutaneous sites: an experimental study. J Craniomaxill Surg 1998;26: I 12-5. [36] Zegzula HD, Buck DC, Brekke J , Wozney JM, Hollinger JO. Bone formation with use of rhBMP-2 (recombinant human bone morphogenetic protein-2). J Bone Joint Surg Am l997;79: 177890.
Journal of' Orthopaedic Research 22 (2004) 514-519 www.elsevier.com/locate/orthres
Long-term effects of local growth factor (IGF-I and TGF-p l ) treatment on fracture healing * A safety study for using growth factors Gerhard Schmidmaier *, Britt Wildemann, Daniel Ostapowicz, Frank Kandziora, Richard Stange, Norbert P. Haas, Michael Raschke Dipirtiiiriif
of' Trrrzollu
ciid
Riwmstnrcriw Sitrgi,rj,, Hurllholr/t Uiiiwrsitj, ( J / Biviin. Charif&,Curllpus Virclioit~C h i c , Augu.strrrh~rrgcrPlirr: I, Bcrliir D- 133.53, Gerillurly
Accepted 26 September 2003
Abstract
Previous studies showed that growth factors dramatically stimulate healing processes in bone. However, the long-term effect of locally applied growth factors on fracture healing remains unclear. In considering the safety of using growth factors, it is necessary to elucidate that after initial stimulation, the effect stops and the result is a normally healed tissue. Therefore, the purpose of the present study was to investigate the long-term time course of healing processes during growth factor (GF) stimulated and unstimulated fracture healing in a closed tibia1 fracture model in rats. A well established local drug delivery system was used. IGF-I (50 pg) and TGF-PI (10 pg) were locally applied using a 10 pin thin polylactide (PDLLA) coating on intramedullary implants. The biomechanical and histomorphometrical results demonstrated a significant stimulation of the fracture healing due to the locally applied growth fxtors compared to control at days 28 and 42 in agreement with the literature. At the last time point, 84 days after fracture, no differences were measurable in the biomechanical tcsting and the callus composition bctween the groups. The callus was consistently in the late phase of remodeling with no remaining cartilage. In conclusion, local growth factor application enhances the healing in the early phase without alteration of the normal healing process. 0 2004 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. Ki,jiwrt/.s:
Long-term effect; Drug delivery; IGF-I: TGF-PI: Fracture healing
Introduction In orthopaedic surgery, great interest continues in efforts to shorten fracture healing time, to increase the rates of unions o r complete regeneration, a n d to treat non-unions successfully. Since the discovery of Marshal Urist in 1965 that demineralized bone can induce ectopic bone formation [30] and that growth factors are responsible for this induction, the research o n this topic has increased dramatically [19]. The number of identified growth factors involved in bone healing processes increased in recent decades. These factors have been characterized [ 151, and the distribution a n d quantifica-
Parts of this study have been presented at the ORS 2003. *Corresponding author. Tel.: +49-30-450-552043: fax: +49-30.45@ 552943. E-riluil ur/c/ws.s: [email protected] (G. Schmidmaier).
tion in the callus tissue [5,8,10,26,34], the mode of action on different cell types [7,11,16,18,29], the signaling cascades [14,12,9] have been analyzed. F o r growth factors such as insulin-like growth factor-I (IGF-I) and transforming growth factor-beta1 (TGF-Pl), a stimulating effect on bone healing was clearly demonstrated in several studies using different animal models [2,4,23,28]. A benefit in using these growth factors is that they d o not induce ectopic bone formation, which has been described for BMPs [35]. Local application of growth factors to enhance bone healing has been accomplished through various carrier systems o r by direct injection. Injection o r local release of TGF-P from tricalcium-phosphate o r hydroxyapatite stimulated bone healing [ 131. Continuous infusion of IGF-I into the arterial supply of the right hind limb of rats affected cortical a n d trabecular bone formation [23]. A critical-size calvarial defect in rats showed improved healing when treated with I G F using a n osmotic
0736-02666 - see front matter 0 2004 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved doi:l0. 1016/j.orlhres.2003,09.0~9
infusion pump [28]. BMP-2 delivered from collagen sponges enhanced fracture healing in a goat tibial fracture model [32] and in an ulnar osteotomy model in the rabbit [6]. These application techniques, however, require the opening of the fracture, direct injections, or the placement of further devices and may have possible side effects due to the use of carriers like bovine collagen ~251. Despite the very promising results of locally applied growth factors on bone healing, concerns due to the long time effects of the factors remains. For example, a misbalance of TGF-PI can have undesirable effects; high local doses can be associated with unresolved inflammation [31]. Studies investigating the side effects of systemically delivered growth factors revealed severe side effects [27,33]. It is unclear if the processes stimulated by locally applied growth factors follow a physiological pattern. A permanent long-term stimulation of bone healing could result in an abnormal bone structure. The normal remodeling of the fracture callus is an important safety issue for clinical application. The hypothesis of the present study was that locally applied growth factors enhance fracture healing in the early phase, but eventually lead to a normal remodeling of the callus resulting in a comparable bony srructure compared to the unstimulated fractures. To investigate this hypothesis, the effects of IGF-I and TGF-PI on different phases of fracture healing were examined. A previous study showed growth factor stimulation resulted in an increased torsional stability compared to the control group and the unfractured contralateral tibia at 28 and 42 days after fracture [20]. However, no data are available concerning longer term effects. In the present study, the progress of rat fracture healing was traced radiologically, and biomechanical torsional testing and histomorphometric analyses were performed after 28, 42 and 84 days.
Materials and methods Five-month-old female Sprague Dawley rats (Harlan-Winkelmann, Borchen, Germany) ( n = 191) were used for the study. After sedation of the animals with Isofluran (Forene') and intraperitoneal anesthesia with a mixture of ketaminhydrochloride (100 mglml) (80 mg/kg body weight) and xylazin 2'%, (12 mg/kg body weight). a closed midshaft fracture of the right tibia was produced using a special fracture device described previously [20]. After closed reposition, the tibiae were intramedullary stabilized with coated versus uncoated titanium Kirschner-wires (1.0 mm diameter, 3.0 cm length). The growth factors recombinant human IGF-I (R&D System, Wiesbaden. Germany) (5% w/w) and recombinant human TGF-Dl (R&D Systems. Wiesbaden, Germany) (1% w/w) were locally applied using poly(D.L-lactide) (Boehringer. Ingelheim, Germany) as a drug carrier coated on the wires. Three groups were investigated: Group 1: implants uncoated. Group 11: implants coated with PDLLA alone. Group 111: implants coated with IGF-I (50 pg) and TGF-Dl (10 pg).
The properties of the PDLLA coating and the release characteristics of the incorporated growth factors were described previously [22]. Briefly, approximately 50% of the incorporated growth factors were released within the first 48 h followed by a sustained release of a further 30% over the next 40 days. The 10 pm thin coating displayed a high degree of mechanical stability on the metallic implant. After healing periods of 28, 42, and 84 days, 20 rats of each group were killed. and the tibiae were prepared for biomechanical testing or histology. The Animal Experimentation Ethics Committee approved all experiments. Throughout the 84 day observation period, radiographs were taken in posterior-anterior and lateral views using Microvision "C Mammography films (Sterling Diagnostic, Newark, D E ) and a Mobilett Plus X-ray unit (Siemens AG, Erlangen. Germany). The bridging of the fracture callus at days 28.42, and 84 was described using following parameters by two independent observers: (a) complete bridging (four cortices were bridged): ( b ) unilateral bridging (1-3 cortices were bridged); (c) n o bridging (no cortex was bridged). For mechanical testing. ten animals were randomly chosen from those killed at each group/time point. and both tibiae were dissected free from soft tissue. The unfractured side served as a control. and data were expressed in percent to the unfractured tibia. After careful dissection of the bones and removal ofthe K-wire, the proximal and distal ends were embedded into two molds with bone cement (Beracryl. Troller, Fullenbach. Switzerland). Each mold was connected to a pivoted axis. Torsion was applied by a material testing machine (Zwick 1455. Ulm, Germany) loading one of the pivoted axes by a lever. The bone was preloaded by an axial force of 5 N, and constant linear displacement (2 mm/min) was applied by the testing machine. The free axis was connected with a load cell (F,,,, = 50 N. HBM. Germany) that determined the torsional force. For histological and histomorphometrical analyses, the other 10 fractured tibiae from each group/time point were fixed for 2 days in 10'%,normal buffered formaldehyde, followed by dehydration in ascending concentrations of ethanol. and embedded undecalcified in methylmethacrylate (Technovit 9 100, Heraeus, Wehrheim, Germany). Longitudinal sections in a sagittal plane were cut at 5 pm with a Leica SM 2500s microtome (Bensheim, Germany) with a 40" stainless steel knife. Von Kossa and a combination of safranin-orange and light-green stains were used. Histomorphometric parameters in the fracture callus were measured using a Leica DM-RB (Bensheim. Germany) microscope and an image analysis system (KS 400. Zeiss, Gottingen. Germany). Parameters were measured at a magnification of 1.6 times, and the following structural indices were calculated: tibial diameter at the fracture gap (Ti' Dm), including Ct . Wi and Ma Dm (mm), bone density of the corticalis. defined as mineralized area/cortical bone area: Md . Ar/Ct B . Ar (I%>): callus area, CI . Ar (mm'); mineralized area of the CI . Ar, Md Ar/CI. Ar (I%,): and cartilage area of the C1.Ar. Cg.ArlC1.Ar ('%I), where Ar=area. B = bone, Cg = cartilage, Ct = corticalis. Dm =diameter, Ma = marrow, Md = Mineralized, and Wi = Width, all based on Parfitt et al. [ 171. and CI = callus. The region of interest (ROI) was defined by the individual tibial diameter (baseline). The callus was divided into the proximal and distal part, and 1.5 baseline lengths were used to define the ROI of proximal and distal callus halves. The total diameter of the callus was included in the ROI. Histomorphometric comparisons between the groups were based on the mineralized volume of the cortices, the area of the whole calluses, and the mineralized and cartilaginous volume of the calluses
PI]. StntlTticciI unulj,sis
The animals were randomized in a blinded manner for radiographic evaluation and for the histological and biomechanical investigations. Data were compared using one way ANOVA for independent samples. The radiological score was analyzed using the %'-test. Both tests were controlled with Bonferroni correction. Interobserver variability for the radiographic evaluation was determined using kappa statistics. Statistical differences were defined at a 95% confidence level. The values are given as mean k standard deviation. SPSS (release 10.0; SPSS Chicago, Illinois) software supported the statistical evaluation.
Results
Eleven animals were excluded from the study due to death during anesthesia or a complex tibia1 fracture at day 0. Differences in the consolidation were seen at 28 and 42 days after fracture, but not at 84 days. The uncoated group I showed a reduced consolidation compared to the PDLLA group I1 and the growth factor group I11 after 28 and 42 days (Fig. 1, Table 1). All groups showed complete consolidation by 84 days. Mechanical testing of tibiae 28 days after fracture caused failure mainly in the callus region in the uncoated and PDLLA-coated groups or adjacent to the fracture callus in the growth factor group. At 42 days, the fractured tibiae of the uncoated group failed mainly in the callus, while those of the PDLLA and growth factor groups failed in the distal or proximal part of the tibia. All three groups failed outside the callus region 84 days after fracture.
The fractures of the uncoated group revealed a significantly ( p < 0.05) lower maximum load and stiffness 28 and 42 days after fracture compared to the other two groups. The growth factor group I11 showed a significantly higher stiffness and load at 28 and 42 days than the PDLLA-coated group. At 84 days, no significant differences were measurable among any of the three groups. The maximum load and the torsional stiffness were approximately 50's) larger than the unfractured tibia (Fig. 2). At 28 and 42 days, the calluses of all three groups were composed of fibroblasts, cartilage cells, and newly formed trabecular bone (Fig. 3). At 28 days, the healing process was in the phase of enchondral ossification in all three groups. By 84 days, the calluses were in the phase of remodeling, and no cartilage was seen in any animal. The histomorphonietry supported the histological evaluation. Control parameters such as the diameter of the tibiae (baseline) and the mineralized area of the cortices showed no significant differences between the
Fig. I . Radiographs (lateral) of right tibiae 28. 42, and 84 days after fracture and intramcdullary stabiliLation with coated versus uncoated titanium K-wires. The progressed consolidation of the fracture i n the growth f x t o r group (IGF-I + TGF-al) is clearly recognizable compared t o the control group. At 84 days after fracture, the fractures of the control and the gt-owtli factor rats were completely consolidated.
Table I Consolidation of the fractures
Complete bridging Incomplete bridging No bridging * p < 0.05
(L2).
IGF-I + TGF-PI
PDLLA
Uncoated 28
42
84
28
42
84
28
42
84
I
3
10
3
5
I0
5'
9.
10
2
-7
0
3
3
0
4
1
0
7
5
0
4
-7
0
I.
0'
0
C. Sclinriiltnuirr et ul. / Journal of Ortlropurdic Reseurch 22 (2004) 514- 51 9
517
groups at any of the time points (Table 2). At the earliest time point, the growth factor treated animals had significantly less cartilage in the callus compared to group I and I1 (Fig. 4a). This difference was also seen 42 days after fracture. Additionally, the callus size was reduced and the mineralized area of the callus was significantly enhanced in the growth factor group I11 compared to group I and I1 (Fig. 4b). At this time point, the PDLLA group also had significantly less cartilage compared to group I. No differences in the callus composition were measurable 84 days after fracture (Table 2). The callus sizes in all three groups were comparable and slightly reduced compared to those found 42 days after fracture. The mineralization was increased in all three groups compared to the 42 day results, and no cartilage was measurable.
Discussion
Fig. 2. Results of torsion testing 28, 42, and 84 days after fracture in comparison to the contralateral unfractured tibiae (in %). At 28 and 42 days, the largest loads (a) and torsional stimnesses (b) were measured in ) by the PDLLA the growth factor group ( I G F - I + T G F - ~ Ifollowed group. No significant differences were found 84 days after fracture.
An enhancement of the healing process due to the application of IGF-I (50 pg) and TGF-PI (10 pg) growth factors delivered locally from a PDLLA-coated implant [22] was measurable 28 and 42 days after fracture as shown before [ ~ o I~. ~ d showed i enhanced ~ ~ con-~ solidation of the fractures after growth factor treatment, and the torsional stiffness and maximum load increased significantly in the growth factor group compared to
Fig. 3. Histological sagittal sections of tibiae 28, 42, and 84 days after fracture stained with Safranin O/light green. The mineralized proportion stained in green, the cartilage in red. A progressive callus remodeling is visible between the growth factor group and the control group.
Table 2 Histomortlhometric results 84 days after fracture
Tibia diameter (base) (mm) Mineralized Adcortex Ar (%I) Periosteal callus Adbase (mm) Mineralized Ar/Ps. CI .Ar (%I) Cartilage Ar/Ps. CI .Ar ('%)
No significant differences between the groups.
Uncoated
PDLLA
IGF-I + TGF-el
2.7 f 0.3 97.2 f 1 .O 3.8f1.7 91.2 f 5.9 0
2.6 ? 0.1 97.52 1.1 2.7 ? 1.O 96.2 2 3.1 0
2.7 f 0.3 97.4 f 0.7 3.7f 1.1 94.0 f 5.1 0
~
~
518
G. Schmidniuicr el uf. I Journul of' Ortltopurriic Reseurch 22 (2004) 514-51 9
Fig. 4. Histomorphometric analyses of the callus composition 28, 42, and 84 days after fracture: (a) the cartilage area in percent t o the total callus area and (b) the mineralized area in percent to the total callus area (where Ar = area, CI =callus. Dm = diameter, Ps = periosteal, and Ti =tibia. * Denotesp < 0.05 between groups at the same time point).
those of the control and the PDLLA-group and of the unfractured contralateral side. The histomorphological and histomorphometrical analyses supported the biomechanical results showing a significantly accelerated callus remodeling in the growth factor group. However, at the longest time point, 84 days after fracture, no differences were found in any of these measures among the three groups. All investigated fractures were completely consolidated. The strength of the fractured bone was consistently higher by approximately SO'%, compared to the unfractured bone. Callus was detectable that was composed only of mineralized tissue without any cartilage. No differences in callus size and mineralization were observed among the groups. This lack of a long-term effect is important for the safety of using growth factors to stimulate bone healing. A concern in using growth factors is that they might result in an enhanced tissue formation without normal remodeling. This study clearly demonstrates that the growth factors affect the healing in the earlier time points, but that at the late time point the fracture is remodeled comparable to the control groups. Other studies investigated the effect of locally applied growth factors in the early and middle phase of healing [3,4,13,32,36]. All these studies showed a stimulating effect of the applied growth factors on the healing. This study reaffirms that fracture healing is accelerated due to the exogenously applied growth factors without alteration of the physiological processes [20]. At the longest
time point, the callus was in the late phase of remodeling. The cartilage observed at the earlier time points was remodeled, and the callus size was significantly reduced compared to the earlier time points. Furthermore, no side effects such as neoplasia or deficient bone production were found. Using a critical size defect, Baltzer et al. were able to demonstrate better healing due to adenoviral transfer of BMP-2 [l]. However, 12 months after surgery and BMP2 application, the biomechanical stability was still increased compared to a control group. Their main focus was on the effect of the growth factors. Several investigations have been focused on the distribution of different factors and receptors during the very early phases of healing [24,26,34]. Our results are in accordance with a study investigating bone closure of scull defects [2]. The treatment of non-healing scull defects with TGF-P1 lead to a stimulation of bone matrix deposition at the earlier investigated time points without alteration of the normal remodeling at the late healing phase. In summary, the investigation of local growth factor application at different time points with different methods revealed an accelerated healing at the early phases (enchondral ossification and the beginning of remodeling). However, at the last time point investigated, the phase of remodeling, no significant differences occurred in torsional stiffness and callus composition with growth factor treatment. The radiological score showed a complete consolidation of the fractures in all investigated animals. These results lead to the conclusion that the local application of growth factors accelerates the healing in the early phase without alteration of the physiological processes and results in a normal remodeling of the callus.
Acknowledgements This study was supported by Deutsche Forschungsgemeinschaft DFG (Schm 1436/1-1).
References [I] Baltzer AW, Lattermann C, Whalen JD. Wooley P, Weiss K. Grimm M, et al. Genetic enhancement of fracture repair: healing of an experimental segmental defect by adenoviral transfer of the BMP-2 gene. Gene Ther 2000;7:73&9. [2] Beck L, Amento E, Xu Y , Deguzman L, Lee W, Nguyen T, et al. TGF-beta1 induces bone closure of skull defects-temporal dynamics of bone formation in defects exposed to rhTGF-betal. J Bone Miner Res 1993;8:75341. [3] Blokhuis TJ, den Boer FC, Bramer JA, Jenner JM, Bakker FC, Patka P, et al. Biomechanical and histological aspects of fracture healing, stimulated with osteogenic protein-I. Biomaterials 2001; 22~725-30.
[4] Blunienfeld I, Srouji S. Lanir Y, Laufer D. Livne E. Enhancement of bone defect healing in old rats by TGF-beta and IGF-I. Exp Gerontol 2002;37:553-65. [S] Bourque WT, Gross M, Hall B K . Expression of four growth factors during fracture repair. Int J Dev Biol 1993;37:573-9. [h] Bouxsein ML. Turek TJ, Blake CA. D'Augusta D. Li X . Stevens M. et al. Recombinant human bone morphogenetic protein-2 accelerates healing in a rabbit ulnar osteotomy model. J Bone Joint Surg Am 2001;83:1219-30. [7] Canalis E. Effect of insulinlike growth factor I on DNA and protein synthesis in cultured rat calvaria. J Clin Invest 1980;66: 709 19. [8] Cho T.J. Gerstenfeld LC, Einhorn TA. Differential temporal expression of members of the transforming growth factor beta superfamily during murine fracture healing. J Bone Miner Res 2002;17:513-20. [9] Claeys I, Siinonet G, Poels J, Van Loy T, Vercammen L, De Loof A. et al. Insulin-related peptides and their conserved signal transduction pathway. Peptides 2002;23:807-16. [lo] Eingartner C , Coerper S, Fritz J, Gaissmaier C , Koveker G, Weise K . Growth factors in distraction osteogenesis. Immuno-histological pattern of TGF-beta1 and IGF-I in human callus induced by distraction osteogenesis. Int Orthop 1999;23:253-9. [ I I ] Hock J. Centrella M, Canalis E. Insulin like growth factor I has independent effects on bone matrix formation and cell replication. Endocrinology 1988:122:25UiO. [I21 Jones JI, Clemmons DR. Insulin-like growth factors and their binding proteins: biological actions. Endocr Rev 1995;16:3-34. [I31 Lind M. Growth factor stimulation of bone healing. Effects on osteoblasts, osteoinies, and implants fixation. Acta Orthop Scand Suppl 1998;283:2-37. [I41 Massague J. TGF-beta signal transduction. AIIIIURev Biochem 1998;67:753-91. [I 51 Massague J. The transforming growth factor-beta family. Annu Rev Cell Biol 1990;6:597-641. (161 Matsumura T. Whelan MC, Li XQ. Trippel SB. Regulation by IGF-I and TGF-beta1 of swarm-rat chondrosarcoma chondrocytes. J Orthop Res 2000:18:3S1-5. [17] Parfitt AM, Drezner MK. Glorieux FH, Kanis JA, Malluche H, Meunier PJ. et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphonietry Nomenclature Committee. J Bone Miner Res 1987; 2:595-610. [IS] Pfeilschifter J. Oechsner M. Naumann A, Gronwald R, Minne H, Ziegler R. Stimulation of bone matrix apposition in vitro by local growth factors: a comparison between Insulin-like growth factor I, platelet-derived growth factor and transforming growth factor beta. Endocrinology 1990;127:69-75. [I91 Schliephake H. Bone growth factors in maxillofacial skeletal reconstruction. Int J Oral Maxillofac Surg 2002;31:469-84. [20] Schmidmaier G, Wildemann B, Bail H, Lucke M, Fuchs T, Stemberger A, et al. Local application of growth factors (insulinlike growth factor-l and transforming growth factor-l31) from a biodegradable poly(o,l.-lactide) coating of osteosynthetic implants accelerates fracture healing in rats. Bone 2001;28:341-50.
[21] Schmidmaier G, Wildeinann B. Cromme F, Kandziora F, Haas NP, Raschke M. BMP-2 coating of titanium implants increases biomechanicdl strength and accelerates bone remodeling in fracture treatment. Bone 2002;6:618-22. [22] Schmidmaier G, Wildemann B, Stemberger A, Haas NP, Raschke M . Biodegradable poly( D , L-lactide) coating of implants for continuous release of growth factors. J Biomed Mater Res Appl Biomat 2001;58:449- 55. [23] Spencer EM, Liu CC, Si EC, Howard GA. In vivo actions of insulin-like growth factor-I (IGF-I) on bone formation and resorption in rats. Bone 1991:12:21 6. [24] Steinbrech DS, Mehrara BJ, Rowe NM, Dudziak ME, Luchs JS. Saadeh PB, et al. Gene expression of TGF-beta, TGF-beta receptor, and extracellular matrix proteins during membranous bone healing in rats. Plast Reconstr Surg 2000;105:2028-38. [25] Takaoka K, Koezuka M, Nakahara H. Telopeptide-depleted bovine skin collagen as a carrier for bone morphogenetic protein. J Orthop Res 1991:9:902 7. [26] Tatsuyama K, Maezawa Y, Baba H, lmamura Y, Fukuda M. Expression of various growth factors for cell proliferation and cytodifferentiation during fracture repair of bone. Eur J Histochem 2000;44:269-78. 1271 Terrell TG, Working PK, Chow CP, Green JD. Pathology of recornbinant human transforming growth factor-beta 1 in rats and rabbits. Int Rev Exp Pathol 1993;34B:43-67. (281 Thaller SR, Dart A, Tesluk H. The effects of insulin-like growth factor- 1 on critical-size calvarial defects in Sprague-Dawley rats, Ann Plast Surg 1993:31:429~33. [29] Tsukazaki T, Usa T, Matsumoto T, Enomoto H, Ohtsuru A. Namba H, et al. Effect of transforming growth factor-beta on the insulin-like growth factor-] autocrinelparacrine axis in cultured rat articular chondrocytes. Exp Cell Res 1994215% 16. [30] Urist M R . Bone formation by autoinduction. Science 1965;ISO: 893-9. [31] Wahl S M . Transforming growth factor beta: the good, the bad, and the ugly. J Exp Med 1994:180:1587-90. [32] Welch R D , Jones AL, Bucholz RW, Reinert CM. Tjia JS, Pierce WA, et al. Effect of recombinant human bone morphogenetic protein-2 on fracture healing in a goat tibia1 fracture model. J Bone Miner Res 1998;13:1483-90. [33] Wilton P. Treatment with recombinant human insulin-like growth factor I of children with growth hormone receptor deficiency (Laron syndrome). Acta Paediatr Suppl 1992:383: 137-42. [34] Yu Y, Yang JL, Chapman-Sheath PJ, Walsh WR. TGF-beta, BMPS, and their signal transducing mediators, Smads, in rat fracture healing. J Biomed Mater Res 2002:60:392 7. [35] Yoshida K, Bessho K, Fujimura K, Kusumoto K. Ogawa Y, Tani Y, et al. Osteoinduction capability of recombinant human bone morphogenetic protein-2 in intramuscular and subcutaneous sites: an experimental study. J Craniomaxill Surg 1998;26: I 12-5. [36] Zegzula HD, Buck DC, Brekke J , Wozney JM, Hollinger JO. Bone formation with use of rhBMP-2 (recombinant human bone morphogenetic protein-2). J Bone Joint Surg Am l997;79: 177890.