J Oral

Maxillofac

57:409419,

Surg 1999

In Vitro Effects of Therapeutic Ultrasound on Cell Proliferation, Protein Synthesis, and Cytokine Production by Human Fibroblasts, Osteoblasts, and Monocytes
Doan, MPH, * Peter Rehem, DDS, MSc, f Sajeda Meghji, BSc, Mphil, PbD,$ and Malcolm Harris, BDS, MD, FDSRCS, FFDRCSIJ
Purpose: The aim of this study was to evaluate several in vitro effects of ultrasound that could revert or prevent the hypoxia, hypovascularity, and hypocellularity observed in osteoradionecrosis. and Methods: Two different ultrasound machines were evaluated, a traditional (1 MHz, pulsed 1:4) and a long wave (45 kHz, continuous) machine, tested at various intensities. Ultrasound was applied to human gingival fibroblasts, mandibular osteoblasts, and monocytes. The assaysperformed were cell proliferation (DNA synthesis), collagen and noncollagenous protein (NCP) synthesis, and cytokine production (ELISA) involving interleukin (IL) lp, IL-6, and IL-S, tumor necrosis factor 01 (TNFa), basic fibroblast growth factor (bFGF), and vascular endothelial growth factor (VEGF). Materials

Nghiem

Both ultrasound machines induced increased cell proliferation in fibroblasts and osteoblasts, between 35% and 52%. The collagen and NCP synthesis were also significantly enhanced to levels up to 112%, the best results being with the 45-kHz machine. The ELISA results showed a slight stimulation of IL-l p by all cell types; there was no difference in IL-6 and TNFol levels. The angiogenesis-relatedcytokines evaluated were significantly stimulated: IL-8 and bFGF production was enhanced in osteoblasts,and VEGF production was stimulated in all three cell types. Both ultrasound machines produced the same results, with the recommended intensities being 15 and 30 mW/cm 2(sA) for the 45-kHz ultrasound, and 0.1 and 0.4 W/cm2(sApA) for the 1 MHz ultrasound.
Results:

Therapeutic ultrasound induces in vitro cell proliferation, collagen/NCP production, bone formation, and angiogenesis. These findings support its use in prospective clinical trials for the prevention and treatment of osteoradionecrosis.
Conclusions:

Osteoradionecrosis (ORN) is a serious long-term complication of radiation therapy involving a triad of hypovascularity, hypocellularity, and hypoxia. l The incidence of ORN of the jaws varies, with reported

ranges from 0.7* to 44.2%.3With adequate prevention, the incidence is still approximately 2% to 5%.4Treatment options for ORN include 1) antibiotics and curettage5Jj; 2) hyperbaric oxygen (HBO) therapy,

*MSc Student, and Maxillofacial UK. tAssistant Horizonte,

Oral and Maxillofacial Surgery, Eastman Federal and Surgery, PhD

Surgery, Dental

Department

of Oral London,

Supported Ministry Address

by a scholarship Brazil. correspondence

(to P.R.) from and reprint Surgery,

CAPES (no. 0469-95/5), to Dr Meghji: WClX Departfor Oral

Institute/UCL, of Minas Department Institute/UCL,

of Education,

requests

Professor, Brazil;

University Student, Dental

Gerais, of

Belo Oral

ment of Oral and Maxillofacial Health Kingdom; Care Science,

Eastman Dental Institute

256 Grays

Inn Rd, London

SLD, United

and Maxillofacial UK.

Eastman

London, Surgery,

e-mail: s.meghji@eastman.ucl.ac.uk
Association of Oral and Max~llofacial Surgeons

o 1999Amer1can

*Senior Lecturer, Department of Oral and Maxillofacial Eastman Dental Institute/lJCL, London, UK. SHead of Department, Surgery, Eastman Dental Department Institute/UCL,

02786386/57040010$3.00/0

of Oral and Maxillofacial London, UK.

409

410
with or without surgery? 3) debridment and local flaps; 4) resection and reconstruction9J0; and recently, 5) therapeutic ultrasound.l Harrisll has suggested the use of ultrasound as an important means of revascularization of mandibular osteoradionecrosis. The patients were treated with ultrasound (3 MHz, pulsed 1:4, 1 W/cm*@~*)) for 40 sessionsof 15 min/d. Ten of 21 (48%) casesshowed healing when treated with debridement and ultrasound alone, 11 casesremained unhealed after ultrasound therapy and after debridement were covered with a local flap, and only one case needed mandibular resection and reconstruction. In contrast, using conventional HBO therapy and surgery for treatment of ORN, Marx8 reported that only 15% of the cases achieved complete healing, and 70% required resection and major reconstruction. It is clear that HBO does not cure all cases of osteoradionecrosis. The use of HBO has also been severely limited by the following factors: 1) very few centers have accessto HBO; 2) the cost of treatment is high; 3) the treatment may be hazardous for the patient; and 4) it is very time consuming.*J1J2 Therefore, the search has continued for a therapy that is practical and accessable. The use of ultrasound appears to be a preferable option. l 1 Ultrasound (US) can be defined as sound wave or pressure wave with a frequency above the limit of the human hearing range (16 to 20 kHz).13 The unit of ultrasound is the Hertz or cycle per second. Being a propagating pressure wave, US is capable of transferring mechanical energy into the tissues.* The energy of the US signal is absorbed, propagated, or reflected, depending on its frequency.15 US can be divided into three types: 1) Diagnostic US uses a frequency between 3 and 5 MHz and low intensity (1 to 50 mW/cm2) to avoid tissue heating; 2) Disruptive ultrasound, such as those used in ultrasonic cleaning devices, usesa very low frequency (20 to 60 kHz), and high intensity above 8 W/cm2; and 3) Therapeutic US, as used in medicine and physiotherapy, usually uses frequencies between 1 and 3 MHz and intensities of 0.1 to 2 .OW/cm2 csApA). Recently, a new US device has been developed that, instead of using the traditional frequencies of 1 to 3 MHz, uses long wave ultrasound at 45 kHz.16 This lower frequency/long wavelength combination gives a widely divergent field shape, with the treated volume effectively in the far field region. This wave penetrates much deeper into the tissues, reaching areasasdeep as several centimeters, instead of millimeters as with the megahertz machines. To minimize heating effects, it uses low intensities (5 to 50 mW/cm2@*)). For clarity, the intensity measurements used here for continuous ultrasound, are spatially averaged intensity (SA), and for pulsed ultrasound are spatial average pulsed averaged (SAPA).

IN VITRO EFFECTS OF THERAPEUTIC

ULTRASOUND

Therapeutic US can exert its physical effects on the cells and tissues by thermal and nonthermal mechanisms. Thermal effects are used in physiotherapy for the treatment of acute injuries, strains, and pain relief. Nonthermal effects are used in the stimulation of tissue regeneration, *,18 healing of varicose ulcers19 and pressure sores,2oblood flow in chronically ischemit muscles,21protein synthesis in fibroblasts,22-24 and tendon repair. 25Ultrasound affects bone by induction of bone formation in vitro26 and acceleration of bone repair in animals14J7-30 and humans.31J2 It has been shown that the nonthermal effects can result in healing of mandibular osteoradionecrosis.l 1 The principal value is the induction of angiogenesis,as shown by Young and Dyson.33 The new capillary formation involves activation, degradation of basement membrane, migration and proliferation of endothelial cells from preexisting venules, capillary tube formation, and maturation of new capillaries.3*Therapeutic angiogenesis is used to reduce unfavorable tissue effects caused by local hypoxia, including osteoradionecrosis, and to enhance tissue repair.s5 The purpose of this study was to clarify the role of therapeutic US in osteoradionecrosis. Cell proliferation assay (DNA synthesis) was used as a model to address the hypocellularity; collagen and noncollagenous protein synthesis assays were used as a model for connective tissueformation, including bone formation; and to address the hypovascularity and hypoxia questions, the production of inflammatory cytokines and known angiogenic factors was evaluated. To determine the most effective ultrasound regimen, two US machines (1 MHz and 45 kHz) were tested at several intensities.

Material

and

Methods

CELL CULTURES

The three cell types usedwere humanmandibularosteoblasts, gingival fibroblasts, and peripheral blood mononuclearcells(monocytes). Fibroblasts and Osteoblasts The osteoblasts were cultured from bone specimens obtainedfrom patientsundergoingsurgicalremovalof third molarsusingthe split bone technique. All patients had no known diseaseand were 20 to 30 years old. Gingival fibroblasts were obtainedin a similarmanner.The bone and gingivalspecimens were rinsedseveral timeswith phosphatebuffered saline (PBS-GIBCO, Grand Island, NY), minced,
and cultured in 75-cm2 cultured flasks using Dulbeccos modified Eagle medium (DMEM) supplemented with: 1) heat-inactivated foetal bovine serum (HIFBS), lO%(v/v) (Sigma, St Louis, MO); 2) freshly prepared L-ascorbic acid, 50 pg/mL (Sigma); 3) L-glutamine, 2 mmol/L (Sigma); and 4) penicillin/streptomycin, 100 U/mL each (Gibco BRL). The flasks were transferred into a humidified 5% C02/95% air incubator at 37C. The media were changed twice a week. After approximately 10 days, the cells started to grow out of the explants. When the cells were confluent, they were trypsinized (0.25% w/v trypsin in PBS) and divided 1 into 3.

DOAN ET AL The fibroblasts were used between the 6th and IOth passages and the osteoblasts between the 4th and 8th passages. The osteoblastic feature was confirmed with alkaline phosphatase staining. The cells were plated in six-well plates (Corning, Acton, MA), at 3 X lo5 cells/well (1.5 X lo5 for the cell proliferation assay) and returned to the incubator. The next day, the media were changed using 5 mL of 2%(v/v) HIFBS, and the cells were left for at least 1 hour before the ultrasound treatment was applied.

411
handset head type is conic and has an effective radiating area of approximately 12.8 cm2. The intensities evaluated were 5, 15, 30, and 50 mW/cm2 @*). Calibration was performed several times during the experiments, but at least one was done before and after a set of assays. The calibration was performed at Orthosonics Ltd., Ashburton, Devon, and was considered satisfactory if the acoustic intensity ranged from 45 to 55 mW/cmz at power 4, with corresponding values at the other settings.

Peripheral Blood Mononuclear Cells (Monocytes)

Monocytes (PMN) were prepared by Ficoll density gradient centrifugation. Monocytes were extracted from thoroughly screened donated human blood obtained from the North London Transfusion Centre. The blood was diluted in equal parts with RPM1 1640 medium (Sigma), and 35 mL of this blood/RPM1 suspension was carefully layered over 15 mL Ficoll-Histopaque 1077 (Sigma). After centrifugation (1,500g for 30 minutes), the mononuclear cell layer was collected in Falcon tubes, washed with RPM1 1640 medium, and centrifuged at 1,500g for 15 minutes. This wash step was repeated and the pellet resuspended in RPM1 1640 medium containing 2% (v/v) HIFBS, L-glutamine, 2 mmol/L (Sigma), and penicillin/streptomycin, 100 U/mL each (Gibco BRL). The PMNs were counted at 5 X lo6 cells/l.5 mL/well in six-well plates. Plates were incubated for 1 to 2 hours to allow the monocytes to adhere, washed once with the prepared medium, and finally each well was filled with 5 mL medium with 2% HIFBS.

THE ULTRASOUND APPLICATIONMODEL

A thermostatically controlled water bath (Electrothermal, London, England), was used to maintain a constant temperature of 37C during the assays. The tank has an internal diameter of 20 cm and a depth of 5 cm and was covered on the inferior and side walls with ultrasound-absorbing rubber. It was filled with distilled, deionized, demineralized water, changed before each experiment. The cells used in the experiments were prepared in six-well culture plates, which had a diameter of 35 mm, and the plate thickness was 1 mm. The plates were placed floating directly over the water surface, taking care not to let any air bubbles form between the plate and the water surface (Fig 1). The transducer was held by a microscope stand, which was placed over a rotating platform/shaker (Edmund Biihler, Germany, model KL2) set to 30 rotations/min. In this way, the transducer was constantly moved while the US was applied, avoiding the production of standing waves. The whole apparatus (water bath, transducer head, and rotating platform) was set up in a sterile air flow cabinet (Microflow Pathfinder; Intermed, Hamshire, United Kingdom). The transducer head was swabbed with 70% isopropyl alcohol BB (Azowipe; Vernon Carus, Preston, United Kingdom), left to dry, and immersed vertically into the culture well, just touching the surface of the medium. Each well of the culture plate had 5 mL medium, and in this way the distance between the transducer head and the cells/bones was approximately 5 to 6 mm. The transducer head from the 45-kHz US machine has a conical shape; therefore, only about 2 to 3 cm2 of the total area of 12.8 cm* was immersed. However, because most of the energy comes through the center of the head, we believe that there was not much energy loss. Insonation was applied to five wells (n = 5) for each intensity. Each well was insonated for 5 minutes, and the control group was treated in the same way, but with the US generator switched off. To concentrate the medium, 2

THE ULTRASOUND MACHINES EVALUATED

The two US machines evaluated were a traditional ultrasound machine that uses a frequency of 1 or 3 MHz, and a long wave machine that uses a frequency of 45 kHz. The traditional US machine was a Therasonic 1032 unit produced by Electra-Medical Supplies (EMS, Oxfordshire, United Kingdom). This apparatus can be set to work with 1 or 3 MHz and can deliver an intensity ranging from 0.1 to 2.0 W/cm2. It also has a pulsing facility and can be set to continuous or pulsed mode, pulsing 1:2, 1:4, or 1:9. The machine has an electronic control panel, a facility to do an electronic check each time it is switched on, and an alert signal if there is no coupling gel or liquid. The handset head has a flat surface and an effective radiating area of approximately 2.0 cm2. The apparatus was set to 1 MHz, pulsed 1:4 (2 ms on and 8 ms o@), and the intensities evaluated were 0.1, 0.4, 0.7, and 1.0 Watts/cm2 (sApA). Several calibrations were performed during the experiments, but at least one was done before and after a set of assays. The calibration was performed at the Department of Medical Physics, University College London. At each calibration, a full electronic checkup was performed, according to the manufacturers manual. The acoustic output power was measured/ calibrated using a precision ultrasound balance (EMS model 67). After setup of the balance and warming up of the US machine, the measurements were taken at 0.2, 0.4,0.6, 0.8, 1.0, and 1.2 Watts/cm*. The calibration was considered in the admissible range if the error in accuracy of the output readings was 10% or less. The long wave US machine (45 kHz) used in this study was a Phys-Assist unit, produced by Orthosonics Ltd (Ashburton, Devon). This apparatus has a fixed frequency of 45 kHz and can deliver an intensity ranging from 5 to 50 mW/cmz GA). It does not need a pulsing facility and works only in continuous mode, which at low intensities does not produce harmful tissue damage. The machine also has an electronic control panel with a liquid crystal display, and a facility to calibrate itself each time it is switched on. The

Ultrasound transducer

Ultrasound rubber

absorb,ng

Thermostatlcalb controlled heater

waer

tank

($20

cm) England

Electrothermal.

FIGURE 1. Schematic drawing of the ultrasound irradiation model used during the assays. The transducer was kept in motion to avoid standing wave formation and inserted into the culture medium above the cells. These were plated into six well plates that were placed floating over the water tank (37C). The whole system was placed in a sterile air flow cabinet, and insonation was applied for 5 minutes for each well.

412
mL was removed immediately after the US treatment, and the plates were cultured for a further 18 hours at 37C in 5% CO*, 95% air.

IN VITRO EFFECTS OF THERAPEUTIC

ULTRASOUND

ELISAASSAYS FOR INTERLEUKIN-1 f3 (IL-1 p), IL-6, IL-8, AND TUMOR NECROSISFACTORa, (TNF-CY)
These measurements were performed for fibroblasts, osteoblasts, and monocytes. The medium from the fibroblasts and osteoblasts was obtained in the previous experiments before radiolabeling. The monocyte medium was obtained in separate experiments. The following antibodies and standards were used: Coating antibodies (diluted in bicarbonate coating buffer, pH 8.2 to 8.3) used were IL-18-immunoaftinity-puritied polyclonal antibodies from sheep anti-IL-lp serum S77/BM, diluted to 2 pg/mL; IL-6-immunoaftinity-purified polyclonal antibodies from goat anti-rh IL-6 serum G15O/BM, diluted to 1 pg/mL; IL-8-immunoaffinity-purified polyclonal antibodies from sheep anti-human IL-S serum S333/BM, diluted to 2 I.lg/mL; tumor necrosis factor alpha (TNFo)-fast protein liquid chromatography (FPLC)-purified monoclonal mouse anti-human TNFol 101-4, diluted to 2 pg/mL. Detecting antibodies (biotinylated immunoaftinity purified antibodies) used were IL-18 goat anti-IL-18 serum G102/BM (diluted l:l,OOO); IL-6-goat anti-rh IL-~ serum GlSO/BM (diluted 1:500); IL-S-sheep anti-human IL-8 serum S333/BM (diluted l:l,OOO); TNFor-biotinylated FPLC-purified polyclonal antibodies from sheep anti-human TNFa serum H/34 or H/91 (diluted 1:200). Standards (human recombinat standards) were IL-1B (IS. 86/680, 1 pg/mL) and IL-6 (I.S. 89/548, 1 ug/mL), at a concentration range of 8,000 to 1.0 pg/mL; IL-8 (NIBSC 89/520, 1 ug/mL), and TNF-cx (NIBSC 87/650), at a concentration range of 10,000 to 1.0 pg/mL.

CELLPROLIFERATION ASSAY (DNA SYNTHESIS)

This assay was performed for fibroblasts and osteoblasts. After the l&hour period following insonation, 1.5 mL of the medium was removed and stored at -70C for the cytokine assays. The cells were radiolabeled with 5-3H thymidine (Radiochemical Centre; Amersham, Buckinghamshire, United Kingdom) to a final concentration of 0.5 pCi/mL in 1.5 mL culture medium. The cells were reincubated for another 6 hours, and the medium was then removed and 1 mL 5% trichloroacetic acid (TCA) was added, stopping the culture, and left at 4C for at least 2 hours. The TCA was then removed and the cells were washed three times with PBS. Then 300 I.IL of sodium hydroxide (NaOH) 0.5 mol/L was added to each well and left for 20 to 30 minutes at 4C. This was removed and transferred to scintillation vials (Minitubes, Hughes and Hughes Ltd, London, England) containing 200 IJL of acetic acid 0.5 mol/L. Scintillation fluid (Unisolve 1; Koch-Light; London, United Kingdom) was added to each tube (4 mL), and radioactivity measured with a Beta counter (Rackbeta, LKB, Wallac, Finland), with external standardization expressed in disintegrations per minute.

COLLAGEN/NON-COLLAGENOUS PROTEIN SYNTHESIS ASSAYS

This assay was also performed on the fibroblasts and osteoblasts. After the l&hour period after insonation, 1.5 mL of the medium was removed for the cytokine assays. The cells were radiolabeled with 5-H proline (Radiochemical Centre, Amersham) to a final concentration of 2 yCi/mL in 1.5 mL of culture medium. The cells were reincubated for another 6 hours, and thereafter 700 ILL of the medium was transferred to Eppendorf tubes containing 700 uL of 10% TCA (final concentration: 5% TCA) and left at 4C for at least 2 hours. The tubes were centrifuged at 4C (2,500g for 30 minutes) to remove unbound isotope and small peptides from the cells, and the supernatant was discarded. The pellets were dissolved in 1 mL of 0.5 mol/L acetic acid containing pepsin (0.5 mg/mL; EC 3.4.4.1, Sigma) and left at 4C overnight (16 hours). In this way, pepsin extracted the collagen by digesting the noncollagenous protein. Rat acidsoluble collagen was added (100 uL of collagen at 5 mg/mL in 0.5 mol/L acetic acid) to act as a carrier for the newly formed collagen. Collagen was precipitated over 3 hours at 4C by the addition of sodium chloride (NaCl) to a final concentration of 5% (w/v) in 0.5 mol/L acetic acid. The tubes were then mixed gently and centrifuged at 4,000 g for 30 minutes at 4C. The supernatant containing the noncollagenous protein (NCP) was stored in scintillation vial inserts (Minitubes, Hughes and Hughes Ltd., England). The pellets were redissolvedm 1 mL of 0.5 mol/L acetic acid, and the collagen was reprecipitated with NaCl, as described, for 2 to 3 hours. The tubes were then centrifuged again at 5,000 to 8,000 g for 30 minutes at 4C and the second supernatant (NCP) was stored in the same scintillation vial used for the first supernatant. The final precipitates of purified collagen were resuspended in 400 I-IL of 0.5 mol/L acetic acid and transferred into another disposable scintillation vial insert. Each scintillation vial was filled with 4 mL scintillation fluid (Unisolve 1, Koch-Light), and the radioactivity was measured as for the proliferation assays.

Technique
Microtiter plates were coated with 100 FL/well of coating antibody, and the plates were incubated overnight at 4C. Unbound coating antibody was removed by washing the plates three times with wash/dilution buffer, pH 7.2 to 7.4 (NaCI, 0.5 mol/L; NaHaP04, 2.5 mmol/L; NaaHP04, 7.5 mmol/L; and Tween 20, 0.1% v/v). Standards of the cytokines and the supernatants to be tested were added to the remaining wells (100 ltL volumes). Plates were incubated for 4 hours (3 hours for IL-8, and 2 hours for IL-6 and TNFo() at room temperature and washed three times with wash/ dilution buffer. Detecting antibody (100 uL) was added to each of the wells and incubated for a further 1 hour at room temperature. Plates were washed three times with wash/ dilution buffer, and 100 FL avidin horseradish peroxidase (Avidin-HRP, Dako Ltd) diluted 1:4,000 in wash-dilution buffer was added into each well. Plates were incubated for 15 minutes at room temperature before washing three times with wash/dilution buffer. Wells were developed with 100 uL of color reagent (0.4 mg orthophenylenediamine and 0.4 ltL of 30% hydrogen peroxide [H202] in 1 mL of 0.1 mol/L citric acid phosphate buffer, pH 5.0) and incubated for 15 to 30 minutes at room temperature. The reaction was terminated by the addition of 150 uL of 1 mol/L sulfuric acid (HaSO& and the absorbance was measured at 492 nm on a Titertek Multiscan spectrophotometer (Flow, Irwing, Scotland). A standard curve was plotted of the absorbance (optical density) versus the concentration of the standards.

BASIC FIBROBLAST GROWTH FACTOR(BFGF)and VASCULARENDOTHELIAL GROWTH FACTOR(VEGF) IMMUNOASSAY

These proteins were assayed using a quantitative sandwich enzyme immunoassay technique (R&D Systems; Abingdon, Oxon, United Kingdom). Briefly, a monoclonal anti-

DOAN ET AL body specific for bFGF or VEGF was precoated on 96well plates. Standards and samples (assayed in duplicates or triplicates) were pipetted into the wells and left to bind to the antibody. After washing away any unbound substances, an enzyme-linked polyclonal antibody specific for bFGF or VEGF was addedto the wells. After a washto remove any unbound antibody-enzyme reagent, a substrate solution wasaddedto the wellsandcolor developedin proportion to the amountof the protein bound in the first step. The color development was stopped and the intensity measured at 570 nm. Standard curveswere obtainedasusual. STATISTICAL ANALYSIS Each cell insonation experiment was repeated at least twice. The numberof observations for controls andfor each intensity evaluated was five (n = 5). The medium was assayed by enzyme-linkeditnmunosorbentassay(ELISA)at least in duplicate. The results obtained were analyzed in Microsoft Excel (Seattle,WA), using ANOVA singlefactor and Studentst-test for unpaired samples. Significance was acceptedat the P < .05 level or higher.

413
ultrasound (Pig 2C), a more uneven distribution occurred, showing an increase of 32% at 5 mW/cm* and of 35% at 30 mW/cm* (P < .05 and P < .Ol). With 1 MHz ultrasound (Fig 2D), again an increase in DNA synthesis was observed at the higher intensities. This was in the order of 34% at 0.7 W/cm2 (P < .Ol) and of 52% at 1.0 W/cm* (P < .OOl).

COLtAGEN/NONCOLtAGENOUS SYNTHESIS

PROTEIN

Results
CELLPROLIFERATION (DNA SYNTHESIS) The cell proliferation assaysshowed an increase in DNA synthesis with both ultrasound machines. The fibroblast group treated with 45 kHz (Pig 2A) showed an increase of 30% and of 43% with 15 and 50 mW/cm*, respectively (P < .Ol). With fibroblasts treated with 1 MHz ultrasound (Fig 2B), the most significant results were an increase of 47% at 0.7 W/cm2 (P < .Ol) and of 37% at 1.0 W/cm2 (P < .05). When the osteoblasts were treated with 45 kHz

The fibroblast group treated with 45 kHz ultrasound (Fig 3A) showed increases in collagen ranging from 37% to 44%, although this was significant only at 15 and 50 mW/cm* (P < .Ol and P < .05). When using l-MHz ultrasound there was a clear tendency for increased collagen production at the lower intensities (Fig 3B), with increases of 48%, 57%, and 52%, at 0.1, 0.4, and 0.7 W/cm2, respectively (P < .Ol, .05, and .Ol, respectively). The collagen/NCP production by osteoblasts was the most significant, because these are the target cells in the repair in osteoradionecrosis. With the 45-kHz ultrasound regimen, the increased production of collagen was significantly higher than with the 1 MHz regimen, of the order of 112%at 30 mW/cm* (P < .05) (Fig 3C). Furthermore, the noncollagenous protein synthesis was significantly increased at all intensities evaluated (P < .Ol) ranging from 59% to 88%. The l-MHz-treated cells also showed increased collagen production at the lower intensities (Fig 3D), with increases in collagen synthesis of 55% and 38% at 0.1

Penentagrd ccmds

FIGURE 2. Cell proliferation assays [incorporation of 3H thymidine into the DNA] show the results for fibroblasts treated with A, 4.5 kHz ultrasound, and 5, 1 MHz ultrasound. The graphs show the results for osteoblasts treated with C, 45 kHz ultrasound and D, 1 MHz ultrasound. Controls received the same treatment, but with the US generator switched off. Values are given as percentages (%) of the controls + SEM. Significance level as compared with controls (sham insonated): *P < .05, **P < .Ol, ***PC ,001.

IN VITRO

EFFECTS

OF THERAPEUTIC

ULTRASOUND

FIGURE 3. Collagen and noncollagenous protein synthesis assays showing the results for fibroblasts treated with A, 4.5 kHz ultrasound and with 5, 1 MHz ultrasound. C, ihe graph (C) shows the results for osteoblasis treated with 45 kHz ultrasound and D, with 1 MHz ultrasound. Controls received same treatment, but with the US generaior switched off. Values are given as percentages 1%) of the controls + SEM. Significance level as compared with controls (sham insonated): *P < .05; **P < .Ol,

***p<.oo1.

and 0.4 W/cm2, respectively (P -=c.05>; the increase in noncollagenous protein synthesis was only significant at 0.4 W/cm2 (P < .05>. CYTOKINE PRODUCTION The induction of IL-lp synthesis by the three cell types was produced by the 1 MHz ultrasound, but it was only induced in monocytes by the 45 KHz ultrasound. However, the levels of stimulation, although significantly different from controls, were still very low for osteoblasts and fibroblasts. Monocytes showed the highest level of IL-lfi synthesis using 45 kHz, followed by osteoblasts with 1 MHz (Fig 4). The production of IL6 was not stimulated in any of the three cell types, and TNFol production in fibroblasts was low and only reached significance with the 5 mW/cm2 and 45 KHz (data not shown). The angiogenesis-related cytokines, IL-S, bFGF, and

VEGF, were significantly stimulated in the following manner: 1) The production of IL-S was enhanced in the osteoblasts treated with both ultrasound machines (Fig 5), but at higher levels with 1 MHz ultrasound. However, IL-S production in monocytes and fibroblasts did not differ from controls; 2) bFGF production was significantly elevated in osteoblasts (Fig 6) with both US machines, and 3) VEGF was significantly stimulated at higher levels than bFGF in all three cell types (Fig 7). Once again, this occurred with both machines. In summary, the angiogenesis-related cytokines, IL-S and bFGF, were significantly stimulated in osteoblasts, and VEGF was significantly stimulated in osteoblasts, fibroblasts, and monocytes. Both ultrasound machines produced significant results, and the best intensities were 15 and 30 mW/cm2csA) with 45 kHz US, and 0.1 and 0.4 W/cm2(sApA) with 1 MHz US.

FIGURE 4. A, IL-1 p production by monocytes stimulated by 5 minutes of 45 kHz continuous ultrasound. Medium of the cells was collected 18 hours after stimulation and assayed by ELISA. 5, It-1 p production by osteoblasts stimulated by 5 minutes of 1 MHz pulsed 1 :4 ultrasound. Medium of the cells was collected 18 hours after stimulation and assayed by ELISA. Controls on A and B were sham-insonated. Bars show mean values + SEM. Significance level as compared with controls (sham insonated]: *P < .05, **P < .Ol, ***p< ,001.

DOAN ET AL
IL8 @@nl, 1200 FIGURE
5. Ii-8 production by osteoblasts stimulated by A, 45 kHz and b B, 1 MHz ultrasound. Medium o Y the cells was collected 18 hours after stimulation and assayed by ELISA. Controls in A and 5 were sham-insonated. Bars show mean values + SEM. Significance level as compared with controls [sham insonated): *P < .05, **p< .Ol, ***p< ,001. ,100 9w *On ,oo 600 500 400

415

Discussion
Wound healing proceeds through a complex series of events involving inflammatory, proliferative, and remodeling phases.s6 The proliferative phase of wound healing consists of rapid fibroblast growth and increased synthesis of collagen/NCP in response to chemotactic factors released during the inflammatory phase. Thus, the release of inflammatory cytokines and the increase in osteoblast and fibroblast proliferation observed in this study simulate the inflammatory and proliferative phases of wound healing. Wound healing is marked by angiogenesis through which ingrowth of capillaries accompanies fibroblast and osteoblast (in the case of bone) proliferation to form granulation tissue. Finally, fibroblasts maintain collagen production, which accumulates during the remodeling phase. Wound healing and tissue regeneration can be impaired by underlying medical factors such as diabetes mellitus, connective tissue diseases,chronic venous insufficiency, cachexia, smoking, previous radiation therapy, cytotoxic therapy, and infection. Compromised angiogenesis is a major reason for delayed healing or nonhealing in most of these cases. In nonhealing wounds such as ORN, a perceived problem is a nonstimulatory level of hypoxia resulting from inadequate perfusion. Moderate levels of hyp-

oxia may facilitate wound healing by inducing the synthesis of collagen precursors and activating macrophages to stimulate angiogenesis. However, with increased levels of hypoxia, a reduction in fibroblast migration and lower collagen synthesis, with impaired hydroxylation of lysine and proline, can be observed.3 Fibroblasts synthesize an intracellular peptide collagen precursor but fail to release it. Maturation and cross-linking of collagen is inversely proportional to the degree of hypoxia but responds to a small increase in oxygen concentration. Therapeutic angiogenesis is the term used to describe the controlled induction or stimulation of neovascularization and neocellularization for the treatment or prevention of pathologic clinical situations characterized by local hypovascularity.35 Healing and tissue regeneration can be improved or accelerated by therapeutic angiogenesis.Traditionally, surgical methods have been used to achieve therapeutic angiogenesis, that is, the transposition of autologous tissues with uncompromised vasculature and high angiogenie potential such as muscle flapss5 as was used originally by Harris. l1 The classic surgical ways of producing therapeutic angiogenesismight be supplemented in the near future by the local application of angiogenic factors and the implantation of autologous capillary endothelial cells cultured ex vivo. Consider-

FIGURE 6. FGFb production by osteoblasts stimulated by A, 45 kHz and b 5, 1 MHz ultrasound. Medium o Y the cells was collected 1 8 hours after stimulation and assayed by ELISA. Controls in A and 6 were sham-insonated. Bars show mean values + SEM. Significance level as compared with controls [sham insonated): *P < .05, **p< .Ol, ***p< ,001.

416

IN VITRO EFFECTS OF THERAPEUTIC

ULTRASOUND

T **

MGF IwhW

135 /-~--------

--.-.. 1C

FIGURE 7. VEGF production after ultrasound stimulation in (A and 5) monocytes and osteoblasts (C and D). The graphs on the left (A, C) show cells stimulated with 45 kHz ultrasound, and the graphs on the right (6, 0) show cells stimulated with 1 MHz ultrasound. Medium of the cells was collected 18 hours after stimulation and assayed by EllSA. Controls were sham-insonated. Bars show mean values + S.E.M. Significance level as compared with controls (sham insonated]: *P < .05, **f < .Ol, ***PC ,001.

able experimental data, as well as some preliminary clinical data, exist to support the usefulnessof angiogenie factors for therapeutic angiogenesis. Ultrasound therapy is the simplest way of producing therapeutic angiogenesid. Young and Dyson33 reported the induction of angiogenesis by US, observed in rat skin lesions. There is considerable clinical evidence to support Young and Dysons work, as shown by the several US effects mentioned in the literature.*1J4J7-s2 Osteoradionecrosis is specifically an area that benefits from therapeutic angiogenesis.The tissues have a complex metabolic/homeostatic deficiency, bordering on ischemic necrosis, and are prone to breakdown, leading to a chronic nonhealing wound. Therefore, the treatment or prevention of this complication aims to restore the mandibular blood supply as well as restore the normal soft tissue and bone homeostasis. The proliferation assays(DNA synthesis) showed that both US machines were able to induce proliferation in fibroblasts and osteoblasts (Fig 2). This is a crucial effect, becauseosteoradionecrosis is a hypocellular tissue and because these assaysare a measure of connective tissue formation. Our data showed an increase around 30% to 50% in sH-thymidine incorporation, which was similar to that observed with HBO used to stimulate fibroblasts cell proliferation.38 However, this also can be interpreted as a deleterious affect, because the cells may be involved in cellular division and not in the production of collagen and other physiologic proteins. This was observed when

the cells were stimulated with the l-MHz machine, when increased proliferation was observed at intensities that caused reduced collagen production (Fig 2D and Fig 3D). The 45-kHz US showed similar observations in fibroblasts; however, with osteoblasts the proliferation was high at the same intensities that increased collagen synthesis (Fig 2C and Fig 3C). The collagen/NCP synthesis assays showed that both ultrasound machines induced collagen production. However, in osteoblaststhe 45-kHz US machine produced much higher increasesin collagen synthesis (up to 112%) than the l-MHz machine (maximum of 55%) (Fig 3A, 30 With the l-MHz machine, this was more evident at lower intensities, both in fibroblasts and in osteoblasts (Fig 3B, 3D). These results were higher than those observed by Harvey et a1,22 who found an increase of both collagen and NCP synthesis, which was intensity dependent, using human skin fibroblasts insonated in suspension and subsequently cultured in vitro. Fibroblasts exposed to continuous US (0.5 W/cm2 @*> showed a 20% increase in collagen secretion, which was increased to 30% when the US was pulsed (0.5 W/cm2 csApA)). Webster et al24also observed smaller increases in protein synthesis by fibroblasts, 29%using a ~-MHZ signal at 0.5 W/cmZ. We showed that 3 MHz pulsed US stimulates bone formation (collagen and NCP production) using a mouse calvarium model, with the best results at 0.1,0.25, and 0.5 W/cm2 c2@. The NCP synthesis was also stimulated, but it could not be correlated to the collagen synthesis in most of the assays.Furthermore, the NCP synthesis

DOAN ET ;u. increase reached significance only in osteoblasts, and mainly with the 45-kHz US (Fig 3C), where all intensities significantly stimulated the cells. This may be an important observation because the NCP contains many cytokines, growth factors, angiogenic factors, and enzymes that may enhance healing and angiogenesis. The intensity of US previously used in the clinical treatment of osteoradionecrosis was relatively high (3 MHz, pulsed 1:4, 1.0 W/cm2(SAPA)). The favorable results observed could therefore be explained in terms of US promoting angiogenesis33 rather than to the effects on collagen protein synthesis, which is higher at lower intensities. This was supported by the use of near-infrared spectroscopy scans of irradiated mandibles,39 which showed higher levels of deoxyhemoglobin concentration in the osteoradionecrotic mandibles of patients after treatment with this ultrasound regimen. This suggests significant improvement in the metabolic activity of the mandibular tissue, probably due to neo-angiogenesis. This finding led us to study the stimulation of cytokine and angiogenic factors. The role of cytokines and angiogenic factors in wound healing and tissues regeneration is now well documented in the literature (Table 1).40-67Cytokines are considered to be intercellular signals that regulate local and systemic inflammatory and anti-inflammatory responses. In comparison with endocrine hormones, cytokines are produced by a variety of cells rather than by a specialized group of cells3s These substances vary widely in origin, molecular weight, chemical composition, and biological activity. Most are only partly characterized.s* Enzymelinked immunosorbent assays (ELISA) have been used extensively for identifying a variety of cytokines, in particular those that are angiogenic factors. The results presented in this report suggest that the

417
healing effect of US on soft tissues, fractures, and osteoradionecrosis may all be the result of the stimulation of angiogenic factors such as IL-S, bFGF, and VEGF. As can be observed in the ELISA results, particularly VEGF, a potent angiogenic factor,43 was significantly stimulated in all three cell types. Mechanical stress can induce a significant increase in the synthesis of IL-l-like factors in deformed osteoblast cultures.@ However, no study on the effect of US on cytokine production has been documented in the literature. It was notable that the production of IL-lp was stimulated at low levels, and this may enhance the production of IL-S, because it has general immunopotentiating activities. However, IL-~ and TNFo were not significantly stimulated, suggesting that the angiogenesis stimulation may be a mechanism distinct from inflammation. These cytokines could have been released into the insonated media either by disruption of the cell membrane or by an increase in either exocytosis or in permeability of the membrane.33 All cell types used in the experiment were counted, and postirradiation viability tests were performed (cell proliferation). These indicated that the osteoblasts, fibroblasts, and monocytes were not lethally damaged. Therefore, it seems more likely that there was a change in either exocytosis or membrane permeability. The exact cellular mechanism underlying the therapeutic action of US is still unknown. From the current literature, Yang et al* proposed several possibilities. First, the compression of microtubules, or cavitation, could produce oscillatory movement of microbubbles and acoustic streaming, and have a direct effect on the permeability of the cell membrane and second messenger adenylate cyclase activity.@ Such changes in ion or protein transport could consequently modify intracellular signals for gene expression.24,69-71 Second, the effects of mechanical pressure at the cell surface could activate the stretch receptor type of cation channel proposed by Sachs,72 and changes in cation concentration also could modify intracellular signals regulating gene expression. Third, the mechanical energy transferred by the US might activate changes in the attachment of the cytoskeleton to the extracellular matrix. Wang et al73 showed in their tensegrity model that the application of mechanical forces to the cytoskeleton affects cell metabolism and gene expression. This was also confirmed by Sandy,@ who showed that mechanical stress can induce a significant increase in the synthesis of IL-l -like factors compared with nonstressed osteoblast cultures. Fourth, electrical currents in bone may be potentiated by exposure to US energy. Investigators have reported increased potentials as a function of US intensity, frequency, and burst pattern.74z75 Finally, a rise in temperature during ultrasonic exposure may have an effect on cell metabo-

Name JJZGF*O*~ TVPF44 PD-ECGF45x46

E&Q?*-49

Endothelial Cell Migration Yes Yes Yes Yes Yes No effect Yes Yes Yes Yes No (inhibits) Yes Yes
tube formation.65

Endothelial Cell Proliferation Yes Yes Yes Yes Yes No effect/yes Yes Yes Yes Yes No (inhibits) No (inhibits) Yes

Angiotropin50~51 Angiogenin52-54 bFGF%-5 *

IL-85659

IGF-160 TGF@ TGF136264 * T~~~6566 * PDGFG7

*These factors also promote

418
lism. The use of low-intensity US reduces tissue heating and also reduces the possibility of cavitation phenomena, that is, the pulsation of gasor vapor-filled voids in a sound field.76 We have shown a maximum temperature rise of 1.8C at 2.0 W/cm2, but no measurable rise was observed at the best stimulatory dose of 0.1 W/cm2(sApA).26 The use of 45-kHz US machines has been shown to produce less heating than the l-MHz machine. Comparing the effects on angiogenic factor production the two US machines (45 kHz and 1 MHz), one concludes that both produced good results, and that each has a particular intensity range where it is more effective. When using 45 kHz, the best angiogenic response should be obtained by using 15 or 30 mW/cm2cSA).With the l-MHz machine, the choice would be 0.1 to 0.4 W/cm2@APA,. These recommendations are in agreement with the studies on cell proliferation and collagen/noncollagenous synthesis, which also showed these intensities as the most effective. There are several advantages of using the long wave ultrasound, such as a higher penetration depth, less heat production, reduced treatment time, and a spherical head applicator giving a bigger effective treatment area. Because its in vitro effects are similar to or even better then those of the l-MHz machine, we recommend it for treatment of mandibular osteoradionecrosis. HBO has been recommended for the treatment of osteoradionecrosis, but it usually is used asadjunctive preparation for resection and reconstruction.1,7 However, treatment with HBO has several disadvantages, not the least being availability and expense. The use of US seems to be superior because it is accessible, quicker, cheaper, and safer. We have shown that US addressesthe main problems of osteoradionecrosis, hypocellularity by the stimulation of cell proliferation, the enhancement of healing and bone formation through the increase of collagen/NCP synthesis, and the hypovascularity and hypoxia through the stimulation of angiogenesis. This work also shows that US most likely produces angiogenesisthrough the production of IL-S, bFGF, and VEGF, all known angiogenic factors. Prospective clinical trials for the treatment and prevention of osteoradionecrosis using lowfrequency (long wave) therapeutic US are indicated.

IN VITRO EFFECTS OF THERAPEUTIC

ULTRASOUND

References
1. Marx RE: Osteoradionecrosis: ology. J Oral Maxillofac Surg 2. Coffin F: The incidence and of the jaws following head 56:851, 1983 3. Rahn AO, Drone JB: Dental treatment of irradiated oral 74:957, 1967 A new concept of its pathophysi41:283, 1983 management of osteoradionecrosis and neck radiotherapy. Br J Radio1 aspects cancer of the problems, care and patient, J Am Dent Assoc

4. Clayman L: Management of dental extractions in irradiated jaws: A protocol without hyperbaric oxygen therapy. J Oral Maxillofac Surg 55:275, 1997 5. Bedwineck JM, Shukovsky LJ, Fletcher GH, et al: Osteonecrosis in patients treated with definitive radiotherapy for squamous cell carcinomas of the oral cavity and naso- and oropharynx. Radiology 119:665, 1976 6. Morrish RB, Chan E, Silverman S, et al: Osteonecrosis inpatients irradiated for head and neck carcinoma. Cancer 47:1980, 1981 7. Mansfield MJ, Sanders DW, Heimbach RD, et al: Hyperbaric oxygen as an adjunct in the treatment of osteoradionecrosis of the mandible. J Oral Surg 39:585, 1981 8. Marx RE: A new concept in the treatment of osteoradionecrosis.JOralMaxillofacSurg41:351, 1983 9. Mainous EG, Hart GB: Osteonecrosis of the mandible: Treatment with hyperbaric oxygen. Arch Otolaryngol101:173,1975 10. Marx RE, Ames JR: The use of hyperbaric oxygen therapy in bony reconstruction of the irradiated and tissue-deficient patient. J Oral Maxillofac Surg 40:412, 1982 11. Harris M: The conservative management of osteoradionecrosis of the mandible with ultrasound therapy. Br J Oral MaxiUofac Swg 30:313,1992 12. Giebfried JW, Lawson W, Biller HF: Complications of hyperbaric oxygen in head and neck disease. Otolaryngol Head Neck Surg 941508, 1986 13. Williams AR: Ultrasound: Biological Effects and Potential Hazards. New York, NY, Academic Press, 1983 14. Yang K-H, Parvizi J, Wang SJ, et al: Exposure to low-intensity ultrasound increases aggrecan gene expression in a rat femur fracture model. J Orthop Res 14:802, 1996 15. Ziskin MC: Applications of ultrasound in medicine: Comparison with other modalities, in Repacholi MH, Grandolfo M, Rindi A (eds): Ultrasound: Medical Applications, Biological Effects and Hazard Potential. New York, NY, Plenum Press, 1987, pp 4961 16. Bradnock B, Law HT, Roscoe K: A quantitative comparative assessment of the immediate response to high frequency ultrasound and low frequency ultrasound (longwave therapy) in the treatment of acute ankle sprains. Physiotherapy 82:78, 1996 17. Dyson M, Pond JB, Joseph J, et ah The stimulation of tissue regeneration by means of ultrasound. Clin Sci 35:273, 1968 18. Dyson M: Role of ultrasound in wound healing, in Kloth LC, McCulloch JM, Feedar JA (eds): Wound Healing: Alternatives in Management. Philadelphia, PA, Davis, 1990, pp 259-285 19. Dyson M, Franks C, Suckling J: Stimulation of healing of varicose ulcers by ultrasound. Ultrasonics 14:232, 1976 20. Paul BJ, LaFratta CW, Dawson AR, et al: Use of ultrasound in the treatment of pressure scores in patients with spinal cord injuries. Arch Phys Med Rehabil41:438, 1960 21. Hogan RDB, Burke KM, Franklin TD: The effect of ultrasound on the microvascular hemo dynamics in skeletal muscle: Effects during ischaemia. Microvasc Res 23:370, 1982 22. Harvey W, Dyson M, Pond JB, et al: The stimulation of protein synthesis in human fibroblasts by therapeutic ultrasound. Rheumatol Rehabil14:237, 1975 23. Reher P, Doan N, Bradnock B, et al: The introduction of long wave ultrasound therapy for osteoradionecrosis. Oral Dis 3:S23, 1997 (abstr) 24. Webster DF, Pond JB, Dyson M, et al: The role of cavitation in the in vitro stimulation of protein synthesis in human fibroblasts by ultrasound. Ultrasound Med Biol4:343, 1978 25. Enwemeka CS, Rodriguez 0, Mendosa S: The biomechanical effects of low-intensity ultrasound on healing tendons. Ultrasound Med Bioll6:8ol, 1990 27. Dyson M, Brookes M: Stimulation of bone repair by ultrasound. Ultrasound Med Biol9:50, 1983 (abstr) 28. PiEa AA, Mont MA, Nasser PR, et al: Non-invasive low intensity pulsed ultrasound accelerates bone healing in the rabbit. J Orthop Trauma 4:246,1990 29. Tsai CL, Chang WH, Liu TK: Preliminary studies of duration of intensity of ultrasonic treatments on fracture repair. Chinese J Physiol35:21, 1992

DOAN

ET AL

419
promotes invasive54. Hu G-F, Riordan JF, Vallee BL: Angiogenin ness of cultured endothelial cells by stimulation of cellassociated proteolytic activities. Proc Nat1 Acad Sci U S A 91:12096,1994 J, Klagsbrun M, Sasse J, et al: A heparin-binding 55. Folkman angiogenic protein-basic fibroblast growth factor-is stored within basement membrane. Am J Path01 130:393, 1988 56. Klagsbrun M: The fibroblast growth factor family: Structural and biological properties. Prog Growth Factor Res 1:207, 1989 of bone 56. Reher P, Flbeshir EI, Harvey W, et al: The stimulation formation ln vitro by therapeutic ultrasound. Ultrasound Med Biol23:1251, 1997 57. Klagsbrun M: Mediators of angiogenesis: The biological significance of basic fibroblast growth factor (bFGF)-heparin and heparan sulfate interactions. Semin Cancer Biol3:81, 1992 stimulates angiogenesis 58. Hu DE, Hori Y, Fan T-PD: Interleukin8 in rats. IntIammation 17:135, 1993 PJ, Kunkel SL, et al: Interleukin-8 as a 59. Koch AE, Polverini macrophagcderived mediator of angiogenesis. Science 258: 1798,1992 60. Grant MB, Mames RN, Fitzgerald C, et al: Insulin-like growth factor I acts as an angiogenic agent in rabbit cornea and retina: Comparative studies with basic fibroblast growth factor. Diabetologia 36:282, 1993 61. Folkman J: Tumor angiogenesis, in Holland JF, Frei E, Bast RC, et al (eds): Cancer Medicine. Philadelphia, PA, Lea & Febiger, 1993, pp 153-170 62. Beck LS, Deguzman L, Lee WP, et al: TGF-beta 1 accelerates wound healing: Reversal of steroid-impaired healing in rats and rabbits. Growth Factors 5:295, 1991 RA, Stone AM, et al: Transforming 63. Phillips GD, Whitehead growth factor beta (TGF-B) stimulation of angiogenesis: An electron microscopy study. J Submicrosc Cytol Path01 25:149, 1993 64. Wahl SM, Hunt DA, Wakefield LM, et al: Transforming growthfactor beta (IGF-beta) induces monocyte chemotaxis and growth factor production. Proc Natl Acad Sci U S A 84:5788, 1987 65. Fan T-PD: Angiogenesis, in Dale MM, Foreman JC, Fan T-PD (eds): Textbook of Immunopharmacology (ed 3). London, England, Blackwell Scientific Publications, 1993, pp 260-268 66. Harris AL: Antiangiogenesis for cancer therapy. Lancet 349:13, 1997 67. Risau W: How blood vessels form. Int J Microcirc 1 l:SS, 1992 68. Sandy JR: Hormonal and mechanical regulation of osteoblast function (PhD Thesis). University of London, England, 1988 pulsed 69. Ryaby JT, Bachner EJ, Bendo JA, et al: Low intensity ultrasound increases calcium incorporation in both differentiating cartilage and bone cell cultures. Tram Orthop Res Sot 14:15, 1989 70. Dyson M: Non-thermal cellular effects of ultrasound. Br J Cancer 45:165, 1982 71. Mortimer AJ, Dyson M: The effect of therapeutic ultrasound on calcium uptake in fibroblasts. Ultrasound Med Biol 14:499, 1988 72. Sachs F: Mechanical transduction by membrane ion channels: A mini review. Mol Cell Biochem 104:57, 1991 73. Wang N, Butler JP, Ingber DE: Mechanotransduction across the cell surface and through the cytoskeleton. Science 260:1124, 1993 74. Behari J, Singh S: Ultrasound propagation in in vivo bone. Ultrasonics 19:87, 1981 75. Duarte LR: The stimulation of bone growth by ultrasound. Arch Orthop Trauma Surg 101:153, 1983 76. Webster DF, Harvey W, Dyson M, et al: The role of ultrasoundinduced cavitation in the in vitro stimulation of collagen synthesis in human fibroblasts. Ultrasonics 18:33, 1980 77. Robertson VJ, Ward AR: Subaqueous ultrasound: 45 KHz and 1 MHz machines compared. Arch Phys Med Rehabil76:569,1995

30. Wang SJ, Lewallen DG, Bolander ME, et al: Low intensity ultrasound treatment increases strength in a rat femoral fracture model. J Orthop Res 12:40, 1994 31. Heckman JD, Ryaby JP, McCabe J, et al: Acceleration of tibial fracture-healing by non-invasive, low intensity pulsed ultrasound. J Bone Joint Surg Am 76:26, 1994 32. Xavier CAM, Duarte LR: Estimulacao ultra-sonica do calo 6sseo: Aplicacao cllnica. Revista Braslleira de Ortopedia 1873, 1983 33. Young SR, Dyson M: The effect of therapeutic ultrasound on angiogenesis. Ultrasound Med Biol 16:261, 1990 34. Jakobsson AE: Induction and inhibition of experimental angiogenesis: A quantitative immunohistochemical and uhrastructural study (PhD Thesis). University of Goteborg, Sweden, Department of Anatomy & Cell Biology and Department of Pathology, 1994 35. Hockel M, Schlenger K, Doctrow S, et al: Therapeutic angiogenesis [Review]. Arch Surg 128:423, 1993 36. Clark RAF: Basics of cutaneous wound repair. J Dermatol Surg Oncol 19:693, 1993 37. Hunt T, Pai M: The effect of varying ambient oxygen tensions on wound metabolism and collagen synthesis. Surg Gynecol Obstet 135:561, 1972 38. Tompach PC, Lew D, Stall JL: Cell response to hyperbaric oxygen treatment. Int J Oral Maxillofac Surg 26:82, 1997 39. Telfah H: Changes in blood flow, tissue metabolism and fat content in the irradiated mandible detected by means of near infrared spectroscopy (NIRS) (MSc Dissertation). University of London, Eastman Dental Institute, 1995 40. Ferraro N, Houck K, Jakeman L, et al: Molecular and biological properties of the vascular endothelial growth factor family of proteins. Endocr Rev 13:18, 1992 41. Jakeman LB, Armanini M, Phillips HS, et al: Developmental expression of binding sites and messenger ribonucleic acid for vascular endothelial growth factor suggests a role for this protein in vasculogenesis and angiogenesis. Endocrinology 133:848, 1993 42. Kim KJ, Li B, Wlner J, et al: Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumour growth in vivo. Nature 362:841, 1993 43. Thomas KA: Vascular endothelial growth factor, a potent and selective angiogenic agent. J Biol Chem 271:603, 1996 44. Connolly DT, Heuvelman DM, Nelson R, et al: Tumor vascular permeability factor stimulates endothelial cell growth and angiogenesis. J Clin Invest 84:1470, 1989 45. Miyazono K, Okabe T, Urabe A, et al: Purification and properties of an endothelial cell growth factor from human platelets. J Biol Chem 262:4098, 1987 46. Miyazono K, Us&i K, Heldin H: Platelet-derived endothelial cell growth factor. Prog Growth Factor Res 3:207, 1991 47. Brown RA, Weiss JB, Tomlinson IW: Angiogenic factor from synovial fluid resembling that from tumors. Lancet 1:682, 1980 48. Odedra R, Weiss JB: Low molecular weight angiogenesis factors. Pharmacol Ther 49: 111, 1991 49. Schor AM, Schor SL, Weiss JB, et al: Stimulation by a lowmolecular-weight angiogenic factor of capillary endothelial cells in culture. Br J Cancer 41:790,1980 50. Hockel M, Sasse J, Wissler JH: Purified monocyte-derived angiogenic substance (angiotropln) stimulates migration, phenotypic changes, and tube formation but not proliferation of capillary endothehal cells in vitro. J Cell Physioll33: 1, 1987 51. Hockel M, Jung HC, Vaupel P, et al: Purified monocyte-derived angiogenic substance (angiotropin) induces controlled angiogenesis associated with regulated tissue proliferation in rabbit skin. J Clin Invest 82:1075, 1988 52. Harper JW, Fox FA, Shapiro R, et al: Mutagenesis of residues flanking Lys-40 enhances the enzymatic activity and reduces the angiogenic potency of angiogenin. Biochemistry 29:7297, 1990 53. Hu G-F, Strydom DJ, Fettfl, et al: Actin is a binding protein for angiogenin. Proc Natl Acad Sci U S A 90:1217, 1993