- Email: [email protected]
Low-intensity pulsed ultrasound increases the fracture callus strength in diabetic BB Wistar rats but does not affect cellular proliferation
Journal of ELSEVIER
Journal of Orthopaedic Research 20 (2002) 587-592
Orthopaedic Research www.elsevier.com/locate/orthres
Low-intensity pulsed ultrasound increases the fracture callus strength in diabetic BB Wistar rats but does not affect cellular proliferation Gregory P. Gebauer, Sheldon S. Lin *, Heather A. Beam, Pedro Vieira, J. Russell Parsons Depurtrnent of Ortlzopaedics, New Jersey Medical School, Uniuersity of Medicine and Dentistry of New Jersey, M S B G 574, 185 South Orcinge Avenue, Newark. NJ 07103-2714, USA
Abstract Type I diabetes mellitus (DM) is associated with impaired fracture healing. Specifically, DM affects multiple phases of fracture healing including early cellular proliferation and late phases resulting in inferior biomechanical properties. Recent studies demonstrated the utility of pulsed low-intensity ultrasound (US) to facilitate fracture healing. The current study evaluated the effects of daily application of US on mid-diaphyseal femoral fractures in DM and non-DM BB Wistar rats. Immunohistochemical staining for PCNA was used to evaluate cellular proliferation at 2, 4, and 7 days post-fracture. In concordance with previous findings, DM fracture callus demonstrated decreased cellular proliferation. Importantly, the application of US did not significantly alter the proliferation in either DM or control groups. However, mechanical testing revealed significantly greater torque to failure and stiffness in US-treated DM versus non-US-treated DM groups at six weeks post-fracture. Despite the inability of US to affect the early proliferative phase of fracture healing, its application clearly results in improved mechanical properties during the late phases of healing. These findings suggest a potential role of US as an adjunct for DM fracture treatment. 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction Type I diabetes mellitus (DM) is an autoimmune disease that disrupts the body’s ability to produce insulin and thus regulate glucose levels. It has been associated with numerous systemic complications, including the retardation of normal fracture healing. Clinical complications include delayed unions, non-unions, and pseudoarthrosis [ 19,22,23]. Normal fracture healing is a complex process that involves sequential recruitment of cells and expression of genes and gene products. The steps can be described in several overlapping phases. Bleeding at the injury site promotes the aggregation of platelets and the formation of a clot. Platelets secrete signaling factors that initiate the proliferative phase during which osteoprogenitor and undifferentiated mesenchymal cells migrate from surrounding areas to the fracture site, proliferate, and
Corresponding author. Tel.: +1-973-972-4408; fax: + 1-973-9725294. E-nznil acldress: [email protected] (S.S. Lin).
eventually differentiate through either osteogenic or chondrogenic pathways. The proliferative phase is a critical step for successful fracture callus formation. Osteoblasts at the site produce intramembranous bone directly while chondroblasts secrete a cartilaginous intermediate to temporarily stabilize the fracture. This cartilage is mineralized and replaced by woven bone through the process of endochondral ossification. Remodeling marks the final phase, resulting in a material with similar mechanical and structure properties as the original bone. Diabetes mellitus has been shown to negatively affect one or more of the phases involved in fracture healing. Experiments in DM BB Wistar rats and Long Evans rats with streptozotocin-induced (STZ) DM demonstrated decreased cellular proliferation in the early phases of DM fracture healing [2,15,18]. In addition, impaired mechanical strength and stiffness have been observed in the later phases [3,6]. Though it is theorized that early disruption in cellular proliferation in DM rats result in weaker fracture callus, this correlation remains unclear. Weaker fracture callus may be the result of changes during all phases of DM fracture healing.
0736-0266/02/$ - see front matter Q 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved PII: S 0 7 3 6 - 0 2 6 6 ( 0 1 ) 0 0 1 3 6 - X
588
G.P. Gebuuer. et at. I fourrid of Ortlzopaeiiic Reserirch 20 (2002) 587-592
Pulsed low-intensity US, a non-invasive form of pulsed mechanical energy, has been shown to accelerate fracture healing [9,10,14]. Animal investigations reported increases in torsional strength and callus size in addition to reduced healing times with US [1,13,24,28]. The rate of non-union significantly decreased using ultrasound as an adjunct [27]. Supporting evidence shows that the rate of fracture healing in humans has been likewise increased [ 10,141. While pulsed low-intensity US has been shown to accelerate normal fracture healing, its effect on DM fracture healing has not been previously investigated. The purpose of this study was to evaluate the influence of US as a fracture-healing adjunct in a DM rat femoral fracture model. Materials and methods Diabetic model
All research protocols were approved by the Institutional Animal Care and Use Committee at University of Medicine and Dentistry of New Jersey. Experiments were conducted on diabetic prone (DP) and diabetic resistant (DR) BB Wistar rats (Biomedical Research Models, Worcester, MA), a strain with a genetic disposition for developing Type I DM in a manner similar to that seen in humans [16]. The rats were fed ad libum. In the DP animals, urine was monitored three times a week for the presence of glycosuria. If glycosuria was present, blood glucose levels were obtained from the tail vein (Accuchex, Roche Diagnostics, Indianapolis, IN). Once these levels reached 250 mg/dl or greater, an insulin releasing palmitic acid implant (140/0 bovine insulin, designed to release insulin at a steady level of 1 U/day for approximately one month; Linplant, Linshin, Canada) was placed subcutaneously into the neck. Blood glucose (BG) values were monitored three times a week, and the amount of implant adjusted to maintain a level of approximately 300 mg/dl. Blood glucose was maintained at this level to simulate a poorly controlled human patient. Additional blood glucose measurements were taken at the time of surgery and at the time of sacrifice. The use of an internal control as an untreated D M group is impossible with this rat strain, because the disease proves fatal 4 7 days after its onset with no exogenous insulin treatment [7]. Non-DM, control animals received a sham implant consisting of palmitic acid. Weights were obtained at the time of implant placement, at the time of sacrifice. and once a week in between.
Ultrasound treatments
The Sonic Accelerated Fracture Healing System (SAFHS, Exogen, Piscataway. NJ), which applied waves of 30 mW/cm2 intensity at a frequency of 1.5 MHz over an area of 3.88 cm2, was used for the US treatments, The transducers were applied at the site of fracture over an aqueous, high viscosity coupling gel, and held in place by medical tape. The rat was placed into a constrained area with the device taped onto the leg. N o anesthetic agent was used in either group. The rats did not fight significantly against the restraints. Animals were monitored during the course of treatments to ensure that the transducers remained in place. US was applied to the animals in the appropriate experimental groups for 20 min per day starting the day after fracture. Cellular prolijeration
This portion of the experiment evaluated the effect of US on the proliferative phase of fracture healing in DM and non-DM BB Wistar animals. Four experimental groups were created, DM animals treated with US (DM, US+), DM animals receiving no US (DM, US-), nonDM control animals receiving US (C, US+) and those receiving no US (C, US-) (n = 5 for all groups). US+ animals received US treatment everyday post fracture, as described above, and were sacrificed at 2,4, and 7 days post-fracture. Upon sacrifice, the femora were removed by disarticulating the knee and hip joints. Excess muscle was removed. Intestine and brain samples were obtained to serve as positive and negative controls, respectively. All samples were fixed in 4% paraformaldehyde, decalcified, and embedded in paraffin. Six-micron sections were cut, prepared, and stained for proliferating cell nuclear antigen (PCNA) analysis, a protein associated with the S phase of an actively dividing cell [5]. The slides were rehydrated and incubated in 3% hydrogen peroxide for 5 min to block any endogenous peroxidase activity. Slides were then inoculated with either a positive monoclonal anti-PCNA antibody (EPOS Clone PCIO, DAKO, Denmark) or a negative control. The EPOS primary PCNA antibody was previously conjugated to a horseradish peroxidase secondary antibody. Phosphate buffered saline was used as a wash in-between each step. A stable 3,3’-diaminobenzidine (DAB) was used as the chromagen. Positively stained cells were counted under 40x magnification in 12 areas (2 groups of 6) centered on the fracture site and located symmetrically on both corteices. Areas 1 and 6 were located adjacent to the cortical bone within the periosteum. These areas represent periosteal callus. Areas 2 and 5 were located on either side of the fracture at the apex of the fracture callus. Areas 3 and 4 were located on either side of the fracture directly adjacent to the cortical bone. These areas (2-5) were selected to represent gap callus. Positively stained cells were counted manually in each area and then added together to determine the total level of proliferation. See Fig. 1 for a schematic diagram of the counting scheme.
Production of the fractures
The femoral fracture model in DM and non-DM BB Wistar rats was originally described by Beam et al. [2]. Two weeks after the onset of DM, fractures were created in the D M rats. The animals were anesthetized using standard anesthetic dosages [ketamine (60 mg/kg), xylazine (8 mg/kg), and a 0.1 mg/kg dose of bupanorphine]. The right hind leg was shaved and washed in Betadine. Using a sterile technique, a medial para-patellar incision was created, and the patella was dislocated laterally. An 18-gauge needle was introduced into the femoral canal through the intercondylar notch, and the canal reamed in retrograde fashion. A 16-gauge teflon cannula was then placed into the canal to stabilize the femur during the fracture procedure. The right femur was fractured by the methods of Bonnarens and Einhorn using a three-point bending device [4].The cannula was replaced with a 1 mm intramedullary Kirschner wire to stabilize the fracture. The patella was reduced, and the incision closed using sutures. The animals were administered 6.5 mg/kg of enofloxacin subcutaneously, and triple antibiotic ointment was placed over the wound. Anterior-posterior and lateral radiographs were taken to ensure transverse, mid-diaphyseal fractures. Animals with fractures not falling within these parameters were excluded from the study. Rats were placed on a warming blanket until they were able to ambulate, at which time they were returned to their normal cages.
Fig. 1 . Schematic view of PCNA counting areas. The number of positively stained cells in each of these 12 areas was counted using 40x magnification. Areas 1 and 6 represented periosteal callus and were located in the elevated periosteum adjacent to the cortical bone. Areas 2-5 represented gap callus. Areas 3 and 4 were located adjacent to the cortical bone on either side of the fracture. Areas 2 and 5 were located on either side of the fracture at the apex of the fracture callus. Measurements were added together to determine the total level of proliferation.
G. P. Gebauer et al. I Journul of Ortlzopaedic Research 20 (2002) 587-592 Meclzunicul properties
The effect of US on the mechanical properties of fracture healing in DM and non-DM BB Wistar rats was also evaluated. Three groups were utilized, DM animals treated with US (DM, US+), DM animals receiving no US (DM, US-), and non-DM, control animals receiving no US (C, US-) (n = 5 for all groups). US+ animals received US treatment everyday post-fracture, as described above, and were sacrificed at six weeks post-fracture. Dhta collected from DM rats with US treatment were compared with results from a larger study investigating mechanical properties of DM rats [3]. At sacrifice, both the fractured and contralateral femora were harvested, wrapped in saline soaked gauze, and stored at -20 "C until the time of testing. Prior to testing, femora were thawed to room temperature and kept moist in a saline solution. The proximal and distal aspects of each femur were embedded in Wood's Metal (Alfa Aesar, Ward Hill, MA). Each end was then placed in the grips of a hydraulically controlled actuator (MTS Systems, Eden Prairie, MN). The bones were tested in torsion until failure at a constant angular displacement rate of 2"/s. Failure was defined as the point at which increases in angular displacement failed to produce an increase in torque. The mechanical data obtained from the fractured limbs were normalized by the values observed in the contralateral limb. Normalizing the data with the contralateral limb accounted for variations in the properties of bone between animals. Stiffness was determined from the slope of the linear regression curve of the torque vs. angular displacement plot prior to the point of fracture. Raw and normalized data are presented in Table 3. Statistical analysis
The means and standard deviations for all results were calculated. Statistical significance was determined for each time point using oneway analysis of variance (ANOVA).
Results Diabetic model
Of the DP rats, over 80% successfully developed DM at an average age of 80 days. Animals failing to show Table 1 Blood glucose levels Blood glucose (mg/dl)
c, usc, us+ DM, USDM, US+
82.18 f 17.77 94.67 5 13.28 340.20 f33.41' 332.91 f27.94'
Blood glucose levels were measured three times a week after the onset of diabetes. Additional measurements were taken at the time of surgery and at the time of sacrafice. The values reported above represent the average of all of these measurements. * Indicates significant difference relative to nondiabetic rats G. i0.05).
589
signs of DM were excluded from the study. The BG level of each animal was monitored regularly throughout the study (see Table 1). Significant differences between average BG values were observed between the DM and non-DM groups (p < 0.0001). Application of US did not alter the average BG levels in either DM or nonDM, controls. Cellular proliferation
In all four groups, the greatest level of cellular proliferation using the PCNA technique was observed four days after fracture (see Table 2 and Fig. 2). Proliferation in both DM groups (DM, US+, and DM, US-) was significantly lower than that of the non-DM (C, US+, C, US-) groups at all three time points, with the greatest variance observed at 4 days (p < 0.0001). No significant change was observed between the DM, US+ and DM, US- group at days 2, 4, and 7, respectively 01 = 0.7266,0.5953,0.6103). Similar findings were observed in C, US+ and C , US- at days 2, 4, and 7, respectively (p = 0.4071,0.3087,0.3424). Mechanical testing
Six weeks post-fracture, the healing femurs of DM, US- rats failed at a torque that was significantly less than that of C , US- rats (p = 0.0003). Healing femurs of DM, US+ animals failed at a torque that was significantly greater than DM, US- animals 0, = 0.0219). This value, however, was still significantly less than C , USanimals 0, = 0.0416). A significant reduction in stiffness was observed in the healing femurs of DM, US- animals when compared to C, US- animals (p = 0.0010). In contrast, US produced a significant increase in stiffness in the femurs of DM, US+ animals compared to DM, US- animals (p= 0.021 1). This value, however, was still significantly less than that observed in C, US-animals 0, = 0.0042) (see Table 3).
Discussion To our knowledge, this is the first study to analyze the effects of US upon cellular proliferation in a DM femur fracture model. The data demonstrated that US
Table 2 Effect of ultrasound on cellular proliferation in diabetic and non-diabetic rats using PCNA
2 Days
c, usc, us+ DM, USDM, US+
8 9 3 . 4 5 111.5 812.4 k 190.5 479.5 & 151.P 447.2 f 141.0'
4 Days
7 Days
2374.0 f319.3 2203.6 f233.5 1012.25 179.6* 928.0 & 288.V
6 1 1 . 4 f 199.7 687.6 f 107.6 327.7 f 101.7" 290.7 f 67.1'
* Indicates significant differences relative to non-diabetic animals both receiving and not receiving ultrasound treatments (p < 0.05).
G. P. GeDtiirer et NI. I
590
JoiiriiciI
of Ortlioptrctlic Resetrrcli 20 (2002) 587-592
Fig. 2. Immunohistological sections of healing fracture calluses stained with PCNA at four days post-fracture observed under 40x magnification. These pictures represent areas of woven bone formation in the periosteal callus. Brownlred coloration indicates proliferating cells. Arrows point to examples of these cells. (A) Diabetic rat not receiving ultrasound; (B) diabetic rat receiving ultrasound; (C) non-diabetic rat not receiving ultrasound. Note: diabetic animals receiving ultrasound appear similar to those not receiving it.
therapy had no effect upon cellular proliferation in either non-DM or DM fracture callus during the early phase of fracture healing. These findings support the belief that the primary effect of US upon fracture healing occurs mainly during the late, non-proliferative phases [8,17,20,21,26,28]. Numerous studies have demonstrated the effectiveness of low-intensity US on non-DM fracture healing [13,24]. Wang et al. [24] reported a 29% increase in the torsional strength of a seven-day-old femur fracture callus with application of US in Sprague Dawley rats. Ito et al. [13] confirmed these findings with comparable results in non-DM animals. To our knowledge, no previous studies have analyzed the effects of US upon the impaired fracture healing process of a DM animal. Recent studies have begun to elucidate the mechanisms by which low intensity US influences fracture healing. Applying US to osteoblasts in an in vitro model, Ryaby et al. [20,21], demonstrated induction of an intracellular calcium influx with resultant increased secondary messenger activity and levels of TGF-P. TGF-P, a known bone growth factor, stimulates osteoblastic activity and increases bone matrix formation. Massaya et al. [17], recently demonstrated increased growth factor levels (PDGF and bFGF) in US treated osteoblasts in vitro. He theorized that US may accelerate fracture healing by inducing various growth factor secretions. Aggregan, a major proteoglycan in cartilage, is present during the formation of the type I1 collagen scaffolding during the enchondral ossification phase and is believed to play a critical role in this process [12,28]. Wu et al. [26], demonstrated an increased aggregan gene expression in chondrocytes following US therapy. Yang et al. [28], noted similar observations in human osteoblasts. These studies imply that US therapy enhances the fracture healing process specifically during the endochondral ossification phase. The mechanical properties of the DM fracture callus have been evaluated in a number of rat fracture models [3,6,11,25]. Wray and Strunkle first observed a decrease in the breaking strength at four weeks of fractured tibiae from DM animals compared to control animals [25]. Unfortunately, these studies neglected to normalize for individual variations in the geometry of the fracture callus. The normalization method described by Funk
Table 3 Mechanical testing results Stiffness (N mddegrees) c , usDM, USDM, US+
* *
37.48 19.09 7.79 6.90* 17.871 5.91"."'
YOof contralateral
Torque (N mm)
'%I of contralateral
83.61 129.96% 26.37 f 11.a%' 42.78 f 7.68'!6*,"
475.97 f 204.72 142.94187.55' 321.17 f 87.21",**
78.18 f 20.25'Yo 26.77 f 11.25%' 55.33 f 20.65%".**
" Indicates significant difference relative to non-diabetic rats not receiving ultrasound treatments (p < 0.05). ** Indicates significant difference relative to diabetic rats not receiving ultrasound treatments @ < 0.05).
59 I
et 211. [6], that takcs into account tlie biomeclianical properties of the intact contralateral bone was used to confirm these eirects in both torque to failure and stiffness testing. Our study demo11st I-atcd that US accelerated fract lire healing in D M animals. The torque to failure of the fracture calluses at six weeks i n tlie US treated D M group was 2.06 times greater (55.33% of contralateral vs. 26.77%) than thc untreated DM group. I n the US treated DM group, stiffiiess values were increased by a factor of 1.62 (42.78% of contralateral vs. 26.37%) comparcd to the DM group not receiving treatment. These results demonstrate that treatment with US leads to increased torque to failure and callus stiffiiess in DM animals. Since US treatmciit failed to affect the initial deficiencies in cellular proliferation observed in DM fracture healing, tlie mecliaiiisin of its action must occur during the late phases of fracture healing. Perhaps the cells in tlie fracture calluses of DM relative to non-DM animals have had their metabolic activity upregulated by the ultrasound treatment. Specific upregulation may be attributed to an increased iiitracellular calcium influx resulting in increased secondary messenger activity and levels of specific bone growth factors (TGF-0) [20,21]. These changes may increase the formation of the cartilagenous soft callus and accelerate the process of endocliondral ossification observed in the late phase of fracture Iiealing. While tlie increase in mechanical strength is perhaps not unexpected, it does show that tlie mecha i i isti1 by w11icli ul t rasouiid increases fracture callus strength is not compromised by diabetes. Our BB Wistar DM rat femur fracture model was previously validated and has been shown to result in impaired fracture Iiea 1i ng. Second a ry to an aut oimmuiie phenomenon, tlie BB Wistar rat develops Type 1 DM during a peripubcrtal age (80 days). If left untreated, liyperglycemia, ketosis, body wasting, and eventually death occur. This model avoids many confounding variables, such as weight loss and cytotoxicity, frequently seen in STZ models, which may influence fracture healing. Our study confirms that the primary effects of US occur during the post-proliferative phase of fracture Iiealing. Despite the inability of US to affect early cellular proliferation. its application clearly results in an iiiiprovement in mechanical properties during tlie late phascs of hcaling. These findings suggest a potential role of US as an adjunct for DM fracture treatment.
Acknowledgciiieiits
This research was funded by 1998 OREF Research Grant (SSL) aiid through tlic Icind donation of the ultrasound ti-catment equipmcnt by Exogcn (Parsippany, NJ).
Refereiices [I] Azuma Y. I t 0 M, Haradii Y. Taknga H,Ohtn T, Jingtishi S. Lowintensity pulse ultrasound accelerates rat femoral fracture healing by ucting on the various cellular reactions i n the fi-acttire callus. J Bone Miner Res 2001:16(4):67 1-80. [2] Beam H, Lin S, Ting S. Dounias C. Ptirsons J . The effect of physiological glticose control on fi'acttire liealing i n diabetic BB Wistar rats. Orthop Res Soc Trans 2000;25:28I . [3] Beam H, Lin S. Satxitino C , Piirsons J . Mechanical ixoperties of . . fractures kind intact bone iii diabetic BB Wistar I'ats. Ortliop Res SOCTrens 2001:26:518. Boiinarens F, Einhorn TA. Production of it standmi fracture in lubofiitory animal bone. J Orthop Res I984;2:97. Fukada K. Morioka H. Imajou S, Ikeda S. Ohtsuka E, Tsurimoto T. Structure function relationship of the etikaryotic DNA replication factor. proliferating cell nuclear antigen. J Biol Chem 1995; 270(38):22527-35. Funk JR, Hale JE. Carmines D. Goocli HL, Hurwitz SR. Biomechanical eveluatioii o f early fracture healing in normal and dkrbetic rats. J Ortliop Res 2000;15:126-32. Gtiberski DL. Diabetes-prone and diabetes-resist~uit BB rats. Animal niodels of spontaneous and vifiilly diebetes mellitus, lyiiiphocytic tyroiditis. and collagen-induced arthritis. ILAR News 1994;35(2):29-37. Hadjiargyrou M. McLeod K, Halsey M, Rubin CT. The tempol-al expression of osteopontin iiiRNA in the fracture callus is altered by low intensity ultrasound. J Bone Minei- Res 1997;12:45. Hadjiargyroti M, McLeod K. Ryaby JP. Rubin C . Enhancement of frecture healing by tiltrasound. Clin Orthop Re1 Res 1998:355S: S2 16-29. Heckman JD, Ryaby JP, McCabe J, Frey JJ. Kilcoyne RF. Acceleration of tibia1 fractLire-healing by non-invasive. low intensity pulsed iiltrasoiind. J Bone Joint Surg [Am] I994;76A: 26-34. Herbsman H, Powers JC, Hirscliman A, Shartan GW. Retardation of fracttlre healing in experimental diabetes. J Surg Res 1968: 5:424-3 I , Hiltunen A. Aro HT, Vuorio E. Regulation ofextracellu1;lr matrix genes during frncture healing in mice. Clin Orthop Re1 Res I993;297:23-7. I t 0 M, Amnia Y . H a r d a Y, Takagi H. Ohta T, Komoriya K, et al. Low intensity pulsed ultrasound accelerates fracture healing in a rat femofiil fracture model. Ortliop Res SOCTrans 1995;44: 732. Kritiansen TR. Ryiiby JP, McCabe J . Frey JJ. Roe LR. Accelerated healing of distal radial fractures with the use of specific, low-intensity ultrasound. J Bone Joint Surg [Am] 1997:79A: 96 I ~-73. Mncey LR, Kana SM. Jinguishi S, Terek RM. Borretos J. Bolander ME. Defects oC early li-;icture liealing in experimentid diabetes. J Bone Joint Siirg [AM] 1989;71:722-33. Marliss EB. Nakhooda AF. Poussier P, Sima Aal: The diabetic syiidroiiie of the BB Wistar rat: possible relevance to type l diabetes in man.Diabetolgia I982:22:225-32. Masaya I , Yosliiaki A, Toniohiro 0. Keiji K . Elrecis ofultliisound on the co-cultured of osteoblasts with endothelinl cells. Trans lnt Soc Frac Repair 1999: l3(4). Ogasowaro A. Y;imaz;iki M. Nak;ijina F. Goto K. Furt~niotoT. Moriya H, et al. Mechanism o f impaired fracture healing i n experiniental diabetes. Orthop Res Soc Trans 1995;23:1011. Papa J, Myerson M. Girard P. Salvage with a]-thodesis i n intractable dinbetic neuropathic anthroplasty of the foot and ankle. .I Bone Joint Surg [Am] I993:75: 1056-66. Ryaby JT. Rachner EJ. Bendo J. Dalton PF. Taiinenhauni S. Pilla AA. Low intensity pulscd ultl.asound increiises calcium
592 incorporation in both cartilage and bone cell cultures. Orthop Res Soc Trans 1989;14:15. [21] Ryaby JT, Matthew J, Dunrte-Alves P. Low intensity pulsed 111trasound affects adenylate cyclase activity and TGF-p syngthesisin osteoblastic cells. Orthop Res SOCTrans 1992;17:590. [22] Stuart MJ. Morrey BF. Arthodesis of the diabetic inellitus neuropithic joint. Clin Orthop Re1 Res 1990;253:209-1 I , [23] Tisdel CL, Marcus RE, Heiple KG. Triple arthrodesis for diabetic paritalar neuropathy. Foot and Ankle Intl 1995;16:332-8. [24] Wang SJ. Lewallen DG, Bolander ME, Chao E, Ilstrup D M , Greenleaf JF. Low intensity ultrasound treatment increases strength in a t-at feinoral fracture model. J Orthop Res 1994312: 40-7.
[25] Wray JB, Stunkle E. The effect of experimental diabetes upon the breaking strength of the healing fracture in the rat. J Surg Res 1965;1 I :479-8 I . [26] Wu C-C, Lewallen DG, Bolander ME. Bronk J, Kinnick R, Greenleaf JF. Exposure to low intensity ultrasound stimulates aggrecan gene expression by cultured chondrocytes. Orthop Res SOCTrans 1996;21:622. [27] Xavier CAM, Duarte LR. Estiniulaci ultra-sonica de callo osseo: applicaca clinica. Rev Brasileira Orthop 1983:18:73-80. [28] Yang K, Parvizi J. Wang S. Lewallen D, Kinnick R, Greenleaf J , et al. Exposure to low-intensity ultrasound increases aggrecan gene expression in a rat femui- fracture model. J Orthop Res 1996; 14:802-9.
Journal of Orthopaedic Research 20 (2002) 587-592
Orthopaedic Research www.elsevier.com/locate/orthres
Low-intensity pulsed ultrasound increases the fracture callus strength in diabetic BB Wistar rats but does not affect cellular proliferation Gregory P. Gebauer, Sheldon S. Lin *, Heather A. Beam, Pedro Vieira, J. Russell Parsons Depurtrnent of Ortlzopaedics, New Jersey Medical School, Uniuersity of Medicine and Dentistry of New Jersey, M S B G 574, 185 South Orcinge Avenue, Newark. NJ 07103-2714, USA
Abstract Type I diabetes mellitus (DM) is associated with impaired fracture healing. Specifically, DM affects multiple phases of fracture healing including early cellular proliferation and late phases resulting in inferior biomechanical properties. Recent studies demonstrated the utility of pulsed low-intensity ultrasound (US) to facilitate fracture healing. The current study evaluated the effects of daily application of US on mid-diaphyseal femoral fractures in DM and non-DM BB Wistar rats. Immunohistochemical staining for PCNA was used to evaluate cellular proliferation at 2, 4, and 7 days post-fracture. In concordance with previous findings, DM fracture callus demonstrated decreased cellular proliferation. Importantly, the application of US did not significantly alter the proliferation in either DM or control groups. However, mechanical testing revealed significantly greater torque to failure and stiffness in US-treated DM versus non-US-treated DM groups at six weeks post-fracture. Despite the inability of US to affect the early proliferative phase of fracture healing, its application clearly results in improved mechanical properties during the late phases of healing. These findings suggest a potential role of US as an adjunct for DM fracture treatment. 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction Type I diabetes mellitus (DM) is an autoimmune disease that disrupts the body’s ability to produce insulin and thus regulate glucose levels. It has been associated with numerous systemic complications, including the retardation of normal fracture healing. Clinical complications include delayed unions, non-unions, and pseudoarthrosis [ 19,22,23]. Normal fracture healing is a complex process that involves sequential recruitment of cells and expression of genes and gene products. The steps can be described in several overlapping phases. Bleeding at the injury site promotes the aggregation of platelets and the formation of a clot. Platelets secrete signaling factors that initiate the proliferative phase during which osteoprogenitor and undifferentiated mesenchymal cells migrate from surrounding areas to the fracture site, proliferate, and
Corresponding author. Tel.: +1-973-972-4408; fax: + 1-973-9725294. E-nznil acldress: [email protected] (S.S. Lin).
eventually differentiate through either osteogenic or chondrogenic pathways. The proliferative phase is a critical step for successful fracture callus formation. Osteoblasts at the site produce intramembranous bone directly while chondroblasts secrete a cartilaginous intermediate to temporarily stabilize the fracture. This cartilage is mineralized and replaced by woven bone through the process of endochondral ossification. Remodeling marks the final phase, resulting in a material with similar mechanical and structure properties as the original bone. Diabetes mellitus has been shown to negatively affect one or more of the phases involved in fracture healing. Experiments in DM BB Wistar rats and Long Evans rats with streptozotocin-induced (STZ) DM demonstrated decreased cellular proliferation in the early phases of DM fracture healing [2,15,18]. In addition, impaired mechanical strength and stiffness have been observed in the later phases [3,6]. Though it is theorized that early disruption in cellular proliferation in DM rats result in weaker fracture callus, this correlation remains unclear. Weaker fracture callus may be the result of changes during all phases of DM fracture healing.
0736-0266/02/$ - see front matter Q 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved PII: S 0 7 3 6 - 0 2 6 6 ( 0 1 ) 0 0 1 3 6 - X
588
G.P. Gebuuer. et at. I fourrid of Ortlzopaeiiic Reserirch 20 (2002) 587-592
Pulsed low-intensity US, a non-invasive form of pulsed mechanical energy, has been shown to accelerate fracture healing [9,10,14]. Animal investigations reported increases in torsional strength and callus size in addition to reduced healing times with US [1,13,24,28]. The rate of non-union significantly decreased using ultrasound as an adjunct [27]. Supporting evidence shows that the rate of fracture healing in humans has been likewise increased [ 10,141. While pulsed low-intensity US has been shown to accelerate normal fracture healing, its effect on DM fracture healing has not been previously investigated. The purpose of this study was to evaluate the influence of US as a fracture-healing adjunct in a DM rat femoral fracture model. Materials and methods Diabetic model
All research protocols were approved by the Institutional Animal Care and Use Committee at University of Medicine and Dentistry of New Jersey. Experiments were conducted on diabetic prone (DP) and diabetic resistant (DR) BB Wistar rats (Biomedical Research Models, Worcester, MA), a strain with a genetic disposition for developing Type I DM in a manner similar to that seen in humans [16]. The rats were fed ad libum. In the DP animals, urine was monitored three times a week for the presence of glycosuria. If glycosuria was present, blood glucose levels were obtained from the tail vein (Accuchex, Roche Diagnostics, Indianapolis, IN). Once these levels reached 250 mg/dl or greater, an insulin releasing palmitic acid implant (140/0 bovine insulin, designed to release insulin at a steady level of 1 U/day for approximately one month; Linplant, Linshin, Canada) was placed subcutaneously into the neck. Blood glucose (BG) values were monitored three times a week, and the amount of implant adjusted to maintain a level of approximately 300 mg/dl. Blood glucose was maintained at this level to simulate a poorly controlled human patient. Additional blood glucose measurements were taken at the time of surgery and at the time of sacrifice. The use of an internal control as an untreated D M group is impossible with this rat strain, because the disease proves fatal 4 7 days after its onset with no exogenous insulin treatment [7]. Non-DM, control animals received a sham implant consisting of palmitic acid. Weights were obtained at the time of implant placement, at the time of sacrifice. and once a week in between.
Ultrasound treatments
The Sonic Accelerated Fracture Healing System (SAFHS, Exogen, Piscataway. NJ), which applied waves of 30 mW/cm2 intensity at a frequency of 1.5 MHz over an area of 3.88 cm2, was used for the US treatments, The transducers were applied at the site of fracture over an aqueous, high viscosity coupling gel, and held in place by medical tape. The rat was placed into a constrained area with the device taped onto the leg. N o anesthetic agent was used in either group. The rats did not fight significantly against the restraints. Animals were monitored during the course of treatments to ensure that the transducers remained in place. US was applied to the animals in the appropriate experimental groups for 20 min per day starting the day after fracture. Cellular prolijeration
This portion of the experiment evaluated the effect of US on the proliferative phase of fracture healing in DM and non-DM BB Wistar animals. Four experimental groups were created, DM animals treated with US (DM, US+), DM animals receiving no US (DM, US-), nonDM control animals receiving US (C, US+) and those receiving no US (C, US-) (n = 5 for all groups). US+ animals received US treatment everyday post fracture, as described above, and were sacrificed at 2,4, and 7 days post-fracture. Upon sacrifice, the femora were removed by disarticulating the knee and hip joints. Excess muscle was removed. Intestine and brain samples were obtained to serve as positive and negative controls, respectively. All samples were fixed in 4% paraformaldehyde, decalcified, and embedded in paraffin. Six-micron sections were cut, prepared, and stained for proliferating cell nuclear antigen (PCNA) analysis, a protein associated with the S phase of an actively dividing cell [5]. The slides were rehydrated and incubated in 3% hydrogen peroxide for 5 min to block any endogenous peroxidase activity. Slides were then inoculated with either a positive monoclonal anti-PCNA antibody (EPOS Clone PCIO, DAKO, Denmark) or a negative control. The EPOS primary PCNA antibody was previously conjugated to a horseradish peroxidase secondary antibody. Phosphate buffered saline was used as a wash in-between each step. A stable 3,3’-diaminobenzidine (DAB) was used as the chromagen. Positively stained cells were counted under 40x magnification in 12 areas (2 groups of 6) centered on the fracture site and located symmetrically on both corteices. Areas 1 and 6 were located adjacent to the cortical bone within the periosteum. These areas represent periosteal callus. Areas 2 and 5 were located on either side of the fracture at the apex of the fracture callus. Areas 3 and 4 were located on either side of the fracture directly adjacent to the cortical bone. These areas (2-5) were selected to represent gap callus. Positively stained cells were counted manually in each area and then added together to determine the total level of proliferation. See Fig. 1 for a schematic diagram of the counting scheme.
Production of the fractures
The femoral fracture model in DM and non-DM BB Wistar rats was originally described by Beam et al. [2]. Two weeks after the onset of DM, fractures were created in the D M rats. The animals were anesthetized using standard anesthetic dosages [ketamine (60 mg/kg), xylazine (8 mg/kg), and a 0.1 mg/kg dose of bupanorphine]. The right hind leg was shaved and washed in Betadine. Using a sterile technique, a medial para-patellar incision was created, and the patella was dislocated laterally. An 18-gauge needle was introduced into the femoral canal through the intercondylar notch, and the canal reamed in retrograde fashion. A 16-gauge teflon cannula was then placed into the canal to stabilize the femur during the fracture procedure. The right femur was fractured by the methods of Bonnarens and Einhorn using a three-point bending device [4].The cannula was replaced with a 1 mm intramedullary Kirschner wire to stabilize the fracture. The patella was reduced, and the incision closed using sutures. The animals were administered 6.5 mg/kg of enofloxacin subcutaneously, and triple antibiotic ointment was placed over the wound. Anterior-posterior and lateral radiographs were taken to ensure transverse, mid-diaphyseal fractures. Animals with fractures not falling within these parameters were excluded from the study. Rats were placed on a warming blanket until they were able to ambulate, at which time they were returned to their normal cages.
Fig. 1 . Schematic view of PCNA counting areas. The number of positively stained cells in each of these 12 areas was counted using 40x magnification. Areas 1 and 6 represented periosteal callus and were located in the elevated periosteum adjacent to the cortical bone. Areas 2-5 represented gap callus. Areas 3 and 4 were located adjacent to the cortical bone on either side of the fracture. Areas 2 and 5 were located on either side of the fracture at the apex of the fracture callus. Measurements were added together to determine the total level of proliferation.
G. P. Gebauer et al. I Journul of Ortlzopaedic Research 20 (2002) 587-592 Meclzunicul properties
The effect of US on the mechanical properties of fracture healing in DM and non-DM BB Wistar rats was also evaluated. Three groups were utilized, DM animals treated with US (DM, US+), DM animals receiving no US (DM, US-), and non-DM, control animals receiving no US (C, US-) (n = 5 for all groups). US+ animals received US treatment everyday post-fracture, as described above, and were sacrificed at six weeks post-fracture. Dhta collected from DM rats with US treatment were compared with results from a larger study investigating mechanical properties of DM rats [3]. At sacrifice, both the fractured and contralateral femora were harvested, wrapped in saline soaked gauze, and stored at -20 "C until the time of testing. Prior to testing, femora were thawed to room temperature and kept moist in a saline solution. The proximal and distal aspects of each femur were embedded in Wood's Metal (Alfa Aesar, Ward Hill, MA). Each end was then placed in the grips of a hydraulically controlled actuator (MTS Systems, Eden Prairie, MN). The bones were tested in torsion until failure at a constant angular displacement rate of 2"/s. Failure was defined as the point at which increases in angular displacement failed to produce an increase in torque. The mechanical data obtained from the fractured limbs were normalized by the values observed in the contralateral limb. Normalizing the data with the contralateral limb accounted for variations in the properties of bone between animals. Stiffness was determined from the slope of the linear regression curve of the torque vs. angular displacement plot prior to the point of fracture. Raw and normalized data are presented in Table 3. Statistical analysis
The means and standard deviations for all results were calculated. Statistical significance was determined for each time point using oneway analysis of variance (ANOVA).
Results Diabetic model
Of the DP rats, over 80% successfully developed DM at an average age of 80 days. Animals failing to show Table 1 Blood glucose levels Blood glucose (mg/dl)
c, usc, us+ DM, USDM, US+
82.18 f 17.77 94.67 5 13.28 340.20 f33.41' 332.91 f27.94'
Blood glucose levels were measured three times a week after the onset of diabetes. Additional measurements were taken at the time of surgery and at the time of sacrafice. The values reported above represent the average of all of these measurements. * Indicates significant difference relative to nondiabetic rats G. i0.05).
589
signs of DM were excluded from the study. The BG level of each animal was monitored regularly throughout the study (see Table 1). Significant differences between average BG values were observed between the DM and non-DM groups (p < 0.0001). Application of US did not alter the average BG levels in either DM or nonDM, controls. Cellular proliferation
In all four groups, the greatest level of cellular proliferation using the PCNA technique was observed four days after fracture (see Table 2 and Fig. 2). Proliferation in both DM groups (DM, US+, and DM, US-) was significantly lower than that of the non-DM (C, US+, C, US-) groups at all three time points, with the greatest variance observed at 4 days (p < 0.0001). No significant change was observed between the DM, US+ and DM, US- group at days 2, 4, and 7, respectively 01 = 0.7266,0.5953,0.6103). Similar findings were observed in C, US+ and C , US- at days 2, 4, and 7, respectively (p = 0.4071,0.3087,0.3424). Mechanical testing
Six weeks post-fracture, the healing femurs of DM, US- rats failed at a torque that was significantly less than that of C , US- rats (p = 0.0003). Healing femurs of DM, US+ animals failed at a torque that was significantly greater than DM, US- animals 0, = 0.0219). This value, however, was still significantly less than C , USanimals 0, = 0.0416). A significant reduction in stiffness was observed in the healing femurs of DM, US- animals when compared to C, US- animals (p = 0.0010). In contrast, US produced a significant increase in stiffness in the femurs of DM, US+ animals compared to DM, US- animals (p= 0.021 1). This value, however, was still significantly less than that observed in C, US-animals 0, = 0.0042) (see Table 3).
Discussion To our knowledge, this is the first study to analyze the effects of US upon cellular proliferation in a DM femur fracture model. The data demonstrated that US
Table 2 Effect of ultrasound on cellular proliferation in diabetic and non-diabetic rats using PCNA
2 Days
c, usc, us+ DM, USDM, US+
8 9 3 . 4 5 111.5 812.4 k 190.5 479.5 & 151.P 447.2 f 141.0'
4 Days
7 Days
2374.0 f319.3 2203.6 f233.5 1012.25 179.6* 928.0 & 288.V
6 1 1 . 4 f 199.7 687.6 f 107.6 327.7 f 101.7" 290.7 f 67.1'
* Indicates significant differences relative to non-diabetic animals both receiving and not receiving ultrasound treatments (p < 0.05).
G. P. GeDtiirer et NI. I
590
JoiiriiciI
of Ortlioptrctlic Resetrrcli 20 (2002) 587-592
Fig. 2. Immunohistological sections of healing fracture calluses stained with PCNA at four days post-fracture observed under 40x magnification. These pictures represent areas of woven bone formation in the periosteal callus. Brownlred coloration indicates proliferating cells. Arrows point to examples of these cells. (A) Diabetic rat not receiving ultrasound; (B) diabetic rat receiving ultrasound; (C) non-diabetic rat not receiving ultrasound. Note: diabetic animals receiving ultrasound appear similar to those not receiving it.
therapy had no effect upon cellular proliferation in either non-DM or DM fracture callus during the early phase of fracture healing. These findings support the belief that the primary effect of US upon fracture healing occurs mainly during the late, non-proliferative phases [8,17,20,21,26,28]. Numerous studies have demonstrated the effectiveness of low-intensity US on non-DM fracture healing [13,24]. Wang et al. [24] reported a 29% increase in the torsional strength of a seven-day-old femur fracture callus with application of US in Sprague Dawley rats. Ito et al. [13] confirmed these findings with comparable results in non-DM animals. To our knowledge, no previous studies have analyzed the effects of US upon the impaired fracture healing process of a DM animal. Recent studies have begun to elucidate the mechanisms by which low intensity US influences fracture healing. Applying US to osteoblasts in an in vitro model, Ryaby et al. [20,21], demonstrated induction of an intracellular calcium influx with resultant increased secondary messenger activity and levels of TGF-P. TGF-P, a known bone growth factor, stimulates osteoblastic activity and increases bone matrix formation. Massaya et al. [17], recently demonstrated increased growth factor levels (PDGF and bFGF) in US treated osteoblasts in vitro. He theorized that US may accelerate fracture healing by inducing various growth factor secretions. Aggregan, a major proteoglycan in cartilage, is present during the formation of the type I1 collagen scaffolding during the enchondral ossification phase and is believed to play a critical role in this process [12,28]. Wu et al. [26], demonstrated an increased aggregan gene expression in chondrocytes following US therapy. Yang et al. [28], noted similar observations in human osteoblasts. These studies imply that US therapy enhances the fracture healing process specifically during the endochondral ossification phase. The mechanical properties of the DM fracture callus have been evaluated in a number of rat fracture models [3,6,11,25]. Wray and Strunkle first observed a decrease in the breaking strength at four weeks of fractured tibiae from DM animals compared to control animals [25]. Unfortunately, these studies neglected to normalize for individual variations in the geometry of the fracture callus. The normalization method described by Funk
Table 3 Mechanical testing results Stiffness (N mddegrees) c , usDM, USDM, US+
* *
37.48 19.09 7.79 6.90* 17.871 5.91"."'
YOof contralateral
Torque (N mm)
'%I of contralateral
83.61 129.96% 26.37 f 11.a%' 42.78 f 7.68'!6*,"
475.97 f 204.72 142.94187.55' 321.17 f 87.21",**
78.18 f 20.25'Yo 26.77 f 11.25%' 55.33 f 20.65%".**
" Indicates significant difference relative to non-diabetic rats not receiving ultrasound treatments (p < 0.05). ** Indicates significant difference relative to diabetic rats not receiving ultrasound treatments @ < 0.05).
59 I
et 211. [6], that takcs into account tlie biomeclianical properties of the intact contralateral bone was used to confirm these eirects in both torque to failure and stiffness testing. Our study demo11st I-atcd that US accelerated fract lire healing in D M animals. The torque to failure of the fracture calluses at six weeks i n tlie US treated D M group was 2.06 times greater (55.33% of contralateral vs. 26.77%) than thc untreated DM group. I n the US treated DM group, stiffiiess values were increased by a factor of 1.62 (42.78% of contralateral vs. 26.37%) comparcd to the DM group not receiving treatment. These results demonstrate that treatment with US leads to increased torque to failure and callus stiffiiess in DM animals. Since US treatmciit failed to affect the initial deficiencies in cellular proliferation observed in DM fracture healing, tlie mecliaiiisin of its action must occur during the late phases of fracture healing. Perhaps the cells in tlie fracture calluses of DM relative to non-DM animals have had their metabolic activity upregulated by the ultrasound treatment. Specific upregulation may be attributed to an increased iiitracellular calcium influx resulting in increased secondary messenger activity and levels of specific bone growth factors (TGF-0) [20,21]. These changes may increase the formation of the cartilagenous soft callus and accelerate the process of endocliondral ossification observed in the late phase of fracture Iiealing. While tlie increase in mechanical strength is perhaps not unexpected, it does show that tlie mecha i i isti1 by w11icli ul t rasouiid increases fracture callus strength is not compromised by diabetes. Our BB Wistar DM rat femur fracture model was previously validated and has been shown to result in impaired fracture Iiea 1i ng. Second a ry to an aut oimmuiie phenomenon, tlie BB Wistar rat develops Type 1 DM during a peripubcrtal age (80 days). If left untreated, liyperglycemia, ketosis, body wasting, and eventually death occur. This model avoids many confounding variables, such as weight loss and cytotoxicity, frequently seen in STZ models, which may influence fracture healing. Our study confirms that the primary effects of US occur during the post-proliferative phase of fracture Iiealing. Despite the inability of US to affect early cellular proliferation. its application clearly results in an iiiiprovement in mechanical properties during tlie late phascs of hcaling. These findings suggest a potential role of US as an adjunct for DM fracture treatment.
Acknowledgciiieiits
This research was funded by 1998 OREF Research Grant (SSL) aiid through tlic Icind donation of the ultrasound ti-catment equipmcnt by Exogcn (Parsippany, NJ).
Refereiices [I] Azuma Y. I t 0 M, Haradii Y. Taknga H,Ohtn T, Jingtishi S. Lowintensity pulse ultrasound accelerates rat femoral fracture healing by ucting on the various cellular reactions i n the fi-acttire callus. J Bone Miner Res 2001:16(4):67 1-80. [2] Beam H, Lin S, Ting S. Dounias C. Ptirsons J . The effect of physiological glticose control on fi'acttire liealing i n diabetic BB Wistar rats. Orthop Res Soc Trans 2000;25:28I . [3] Beam H, Lin S. Satxitino C , Piirsons J . Mechanical ixoperties of . . fractures kind intact bone iii diabetic BB Wistar I'ats. Ortliop Res SOCTrens 2001:26:518. Boiinarens F, Einhorn TA. Production of it standmi fracture in lubofiitory animal bone. J Orthop Res I984;2:97. Fukada K. Morioka H. Imajou S, Ikeda S. Ohtsuka E, Tsurimoto T. Structure function relationship of the etikaryotic DNA replication factor. proliferating cell nuclear antigen. J Biol Chem 1995; 270(38):22527-35. Funk JR, Hale JE. Carmines D. Goocli HL, Hurwitz SR. Biomechanical eveluatioii o f early fracture healing in normal and dkrbetic rats. J Ortliop Res 2000;15:126-32. Gtiberski DL. Diabetes-prone and diabetes-resist~uit BB rats. Animal niodels of spontaneous and vifiilly diebetes mellitus, lyiiiphocytic tyroiditis. and collagen-induced arthritis. ILAR News 1994;35(2):29-37. Hadjiargyrou M. McLeod K, Halsey M, Rubin CT. The tempol-al expression of osteopontin iiiRNA in the fracture callus is altered by low intensity ultrasound. J Bone Minei- Res 1997;12:45. Hadjiargyroti M, McLeod K. Ryaby JP. Rubin C . Enhancement of frecture healing by tiltrasound. Clin Orthop Re1 Res 1998:355S: S2 16-29. Heckman JD, Ryaby JP, McCabe J, Frey JJ. Kilcoyne RF. Acceleration of tibia1 fractLire-healing by non-invasive. low intensity pulsed iiltrasoiind. J Bone Joint Surg [Am] I994;76A: 26-34. Herbsman H, Powers JC, Hirscliman A, Shartan GW. Retardation of fracttlre healing in experimental diabetes. J Surg Res 1968: 5:424-3 I , Hiltunen A. Aro HT, Vuorio E. Regulation ofextracellu1;lr matrix genes during frncture healing in mice. Clin Orthop Re1 Res I993;297:23-7. I t 0 M, Amnia Y . H a r d a Y, Takagi H. Ohta T, Komoriya K, et al. Low intensity pulsed ultrasound accelerates fracture healing in a rat femofiil fracture model. Ortliop Res SOCTrans 1995;44: 732. Kritiansen TR. Ryiiby JP, McCabe J . Frey JJ. Roe LR. Accelerated healing of distal radial fractures with the use of specific, low-intensity ultrasound. J Bone Joint Surg [Am] 1997:79A: 96 I ~-73. Mncey LR, Kana SM. Jinguishi S, Terek RM. Borretos J. Bolander ME. Defects oC early li-;icture liealing in experimentid diabetes. J Bone Joint Siirg [AM] 1989;71:722-33. Marliss EB. Nakhooda AF. Poussier P, Sima Aal: The diabetic syiidroiiie of the BB Wistar rat: possible relevance to type l diabetes in man.Diabetolgia I982:22:225-32. Masaya I , Yosliiaki A, Toniohiro 0. Keiji K . Elrecis ofultliisound on the co-cultured of osteoblasts with endothelinl cells. Trans lnt Soc Frac Repair 1999: l3(4). Ogasowaro A. Y;imaz;iki M. Nak;ijina F. Goto K. Furt~niotoT. Moriya H, et al. Mechanism o f impaired fracture healing i n experiniental diabetes. Orthop Res Soc Trans 1995;23:1011. Papa J, Myerson M. Girard P. Salvage with a]-thodesis i n intractable dinbetic neuropathic anthroplasty of the foot and ankle. .I Bone Joint Surg [Am] I993:75: 1056-66. Ryaby JT. Rachner EJ. Bendo J. Dalton PF. Taiinenhauni S. Pilla AA. Low intensity pulscd ultl.asound increiises calcium
592 incorporation in both cartilage and bone cell cultures. Orthop Res Soc Trans 1989;14:15. [21] Ryaby JT, Matthew J, Dunrte-Alves P. Low intensity pulsed 111trasound affects adenylate cyclase activity and TGF-p syngthesisin osteoblastic cells. Orthop Res SOCTrans 1992;17:590. [22] Stuart MJ. Morrey BF. Arthodesis of the diabetic inellitus neuropithic joint. Clin Orthop Re1 Res 1990;253:209-1 I , [23] Tisdel CL, Marcus RE, Heiple KG. Triple arthrodesis for diabetic paritalar neuropathy. Foot and Ankle Intl 1995;16:332-8. [24] Wang SJ. Lewallen DG, Bolander ME, Chao E, Ilstrup D M , Greenleaf JF. Low intensity ultrasound treatment increases strength in a t-at feinoral fracture model. J Orthop Res 1994312: 40-7.
[25] Wray JB, Stunkle E. The effect of experimental diabetes upon the breaking strength of the healing fracture in the rat. J Surg Res 1965;1 I :479-8 I . [26] Wu C-C, Lewallen DG, Bolander ME. Bronk J, Kinnick R, Greenleaf JF. Exposure to low intensity ultrasound stimulates aggrecan gene expression by cultured chondrocytes. Orthop Res SOCTrans 1996;21:622. [27] Xavier CAM, Duarte LR. Estiniulaci ultra-sonica de callo osseo: applicaca clinica. Rev Brasileira Orthop 1983:18:73-80. [28] Yang K, Parvizi J. Wang S. Lewallen D, Kinnick R, Greenleaf J , et al. Exposure to low-intensity ultrasound increases aggrecan gene expression in a rat femui- fracture model. J Orthop Res 1996; 14:802-9.