Effects of 35 kHz, Low-frequency Ultrasound Application In Vitro on Human Fibroblast Morphology and Migration Patterns

Low-frequency ultrasound (LFU) in the frequency range 30–40 kHz administered using different delivery methods (waterbath and noncontact spray) has shown positive effects on chronic wound healing rates in humans, but the underlying mechanisms are only beginning to be explored. To examine the effects of LFU delivered at 35 kHz on the morphology and migration of human fibroblasts, real-time videography was used to record the rate and character of cultured human fibroblast migration at 8-hour increments for 48 hours in a wound assay model.

  Cells were treated with 35 kHz LFU or saline only (control). Cellular morphology (cell shape, packing, and apparent length) and migration patterns including orientation were analyzed using time-lapse videography. LFU delivered at a frequency of 35 kHz produced a different pattern of fibroblast migration in vitro (varied orientation versus vertical orientation for untreated cells) and altered cell morphology compared to controls. The observed pattern of migration was more varied and widely distributed across multiple angles versus the predominant parallel orientation of the migrating untreated cells. The potential implications of these findings on collagen placem

Therapeutic ultrasound has been used to treat a variety of soft tissue injuries for more than 70 years in the United States.^1,2 During this time, research has demonstrated clinically significant effects of ultrasound on dermal healing rates of chronic venous ulcerations.^3-5 The earliest work examined the effects of high-frequency ultrasound (HFU; >1 MHz or 1,000,000 cycles per second) on venous ulcers.³ In a small controlled trial, Dyson et al³ demonstrated increased wound healing in varicose ulcers exposed to 3 MHz ultrasound 3 times per week using athermal parameters of 1 W/cm² and 20% pulsed mode to the periwound area. In controlled clinical trials, Callam et al⁴ and Roche and West⁵ demonstrated similar findings with periwound treatment using high-frequency, low-intensity pulsed ultrasound. The average decrease in wound size for the 3 studies was 17% (27% for Dyson et al,³ 18% for Callam et al,⁴ and 7% for Roche and West⁵). A recent randomized, controlled trial also demonstrated slightly increased healing rates in nonsurgical wounds treated with HFU⁶ compared to a variety of other treatment modalities including compression, electrical stimulation, low-level laser, compression, and these therapies in combination with drug therapy.

  In 1997, the effects of low-frequency ultrasound (LFU) (<1 MHz or 1,000,000 cycles per second) were examined by 2 independent labs.^7,8 Both lab groups tested 30 kHz (30,000 cycles/second) generators set at 100 mW/cm² using a waterbath technique. Treatments were administered 3 times per week for 10 minutes. Both the wound surface and periwound area of venous ulcerations were treated. Weichenthal et al⁸ detected a 30% decrease in wound size in 8 weeks compared to control, and Peschen et al⁷ noted a 38% decrease in wound size in 12 weeks. Likewise, results from a retrospective study⁹ demonstrated a greater percentage of chronic lower extremity ulcers closed in 163 individuals treated with a noncontact, low-frequency 40 kHz ultrasound (53%) generator, compared to 32% of ulcers in 47 individuals who received standard care (ie, advanced wound dressings, debridement, and interventions indicated by etiology).

  In an attempt to elucidate the mechanism through which HFU stimulates healing in chronic venous ulcerations, a number of investigators have examined the effects of this physical energy modality in vitro on fibroblast function. Harvey et al¹⁰ demonstrated that cultured human fibroblasts treated with HFU delivered at 3 MHz and at an intensity of 0.5–2 W/cm² exhibited enhanced protein synthesis. A later study by Ramirez et al¹¹ provided evidence that HFU at 1 MHz stimulated collagen synthesis and fibroblast proliferation when delivered at an intensity of 0.4 W/cm². Additionally, Zhou et al¹² found high-frequency, low-intensity ultrasound facilitated DNA synthesis in fibroblasts, and fibroblast numbers significantly (P <0.05) increased post-treatment. These changes were obtained with a daily treatment paradigm that reflects clinically relevant parameters.

  The importance of calcium in modulating biological process critical to wound healing is well known. In vitro analysis by Pires-Oliveira et al¹³ indicates HFU stimulates the uptake of calcium in fibroblasts; this elevation of cytosolic calcium is time-limited with the resequestration of calcium in the endoplasmic reticulum after cessation of ultrasound exposure. Interestingly, Pires-Oliveira et al¹³ also demonstrated morphological changes in the endoplasmic reticulum after fibroblast stimulation with high-frequency, low-intensity ultrasound levels of 0.2–0.6 W/cm², consistent with calcium resequestration. Furthermore, the elevation of cytosolic calcium is congruent with the findings of several other labs.^10,11

  To date, only 1 scholarly, peer-reviewed article has examined the effects of LFU on fibroblast function. In 2007, Lai and Pittelkow¹⁴ studied the cellular morphology, mitogenic activities, expression of common growth factors involved in wound healing (keratinocyte growth factor [KGF] and transforming growth factor beta-1 [TGF-beta1]), and activation of extracellular signal regulated kinase (ERK) and c-Jun N-terminal kinase (JNK) signaling pathways of dermal fibroblasts post-treatment with 40 kHz noncontact LFU as compared to untreated fibroblasts. The only morphological change detected in the LFU-treated fibroblast was vacuolization. However, mitogenic activity was increased with a much earlier increase in KGF expression, ERK activation, and JNK activation. The ratio of ERK/JNK also was increased in ultrasound-treated fibroblasts.

  The purpose of this in vitro study was to further examine the effects of LFU on human fibroblast morphology and function using real-time videography to record the rate and character of fibroblast migration in 35 kHz LFU-treated and control fibroblasts.

Equipment. The LFU generator used in this experiment operates at a frequency of 35 kHz, with an adjustable treatment intensity of up to approximately 2.0 W/cm² (Arobella Medical LLC, Qoustic Wound Therapy System™ Minnetoka, MN). The treatment delivered in this experiment utilized 100% power output at an intensity of 2.0 W/cm². The intensity was calibrated using a proprietary laser interferometer that characterizes the displacement, frequency, and frequency waveform of the Qoustic Qurette™ applicator tip. In addition, acoustic output pressure, intensity, and power parameters were measured in an anechoic chamber and/or hydrophone tank to confirm accomplishment of desired output characterization (ie, intensity and power versus distance and acoustic field shape).

  Saline jet irrigation was delivered through an orifice inside a metal curette capable of producing a distal displacement from 0–75 µm depending on the percent power setting. The LFU energy was produced through the transduction of 60-cycle wall current into mechanical energy via a titanium alloy transducer using piezoelectric elements made of lead zirconate titanate.

  Mechanical energy was produced as the transducer resonated to create an axial oscillation of the curette that radiates ultrasonically both through the scoop shape as well as off the distal end of the curette. This dual action acts to fragment tissue as well as focus ultrasound energy propagated toward the treatment area through the saline jet via the scoop shape. Sterile normal saline was connected to a port on the hand piece and exited the curette as a saline jet, which serves both to irrigate and cleanse the wound of tissue fragments, debris, exudate, and other matter as well as to provide a coupling medium for the transmission of ultrasonic energy directly or through the scoop shape.

  The hand piece, with its treatment curette probe, was used to deliver LFU in a zigzag fashion across the fibroblast cultures at a calibrated distance by placing the probe directly adjacent to the lip of the culture dish.

  Cell culture. Normal adult human dermal fibroblasts (HDFa; American Type Culture Collection [ATCC], Manassas, VA) were cultured in complete fibroblast basal medium (FBM) comprised of FBM (ATCC) supplemented with 10% fetal bovine serum (Atlanta Biologicals, Inc, Lawrenceville, GA), GlutaMAX ITM (2 mM) (Life Technologies, Grand Island, NY); penicillin and strepytomycin (Sigma Aldrich, St. Louis, MO), 100 U/mL and 100 µg/mL, respectively; and phenol red (2 µM) (Sigma). All cultures were maintained in a humidified 5% CO₂ incubator at 37˚ C.

  Scratch wound assay. HDFa were seeded in 35-mm tissue culture dishes and grown until approximately 90% confluence, so cells were elongated without contact inhibition at the time of the scratch wound assay. By using a model of subconfluence, singular fibroblastic migration patterns could be observed.¹⁵

  All cultures were scratched with a sterile 200-µL plastic pipette tip. A smooth, straight scratch was made through the monolayer of cells by holding the culture dish at a 45˚ angle and allowing the weight of the pipette to rest on the back of the hand. Cells then were treated with 35 kHz LFU or saline only (control). Cells treated with the LFU were sprayed with saline in a slow zigzag sweeping motion for 10 seconds. Saline was removed from both ultrasonic-treated and control cells by aspirating with a pipette. Complete FBM (1 mL) was added to the cells in a slow drip down the side of the culture dish. Culture dishes were positioned on an Olympus Ix70 inverted or a Zeiss Axiovert inverted microscope and magnified 10X. Additional complete FBM (1 mL) was added to the culture dish, and the cells were maintained at 37˚ C and 5% CO₂. Scratch images were taken at approximately 9-second intervals during a 48-hour period.

  Two observers blinded to study or control cultures viewed the images taken at 8-hour increments. For each image collected at 8, 16, 24, 32, 40, and 48 hours, the number, length, and angle of the HDFa cells entering the scratch wound were recorded. Length was determined by measuring the distance between the 2 points at maximal distance on the outer edge of the cell. Average length of the cells entering the scratch wound for each culture dish was calculated and reported.

  Morphological observations. Images taken at 8, 16, 24, 32, 40, and 48 hours using the Olympus Ix70 inverted or a Zeiss Axiovert inverted microscope and magnified 10X were analyzed for morphological changes including cell shape, packing, and apparent length.

  Data collection and statistical analysis. A 2-tailed, paired t-test was used to compare migration rates of treated and control fibroblasts (untreated groups). Comparison of the angulation of fibroblasts entering the artificially created wounds was performed using a 2-tailed, Fisher’s exact test. All graphs were created and statistical analyses were performed using GraphPad Prism 6.00 for Windows (GraphPad Software, Inc, San Diego, CA). Observational analysis of fibroblast morphological changes was performed using the microscopic videography images collected for the migration pattern analysis.

  To compare the rate of closure for untreated and LFU-treated scratch wound assays, the average width of a scratch wound remaining at 3 locations for each wound at 0, 8, 16, and 24 hours was recorded. The percentage of width of a scratch wound remaining at 8, 16, and 24 hours for each wound was calculated. The percentages of width of scratch wounds remaining for untreated and LFU-treated cells were compared at 8, 16, and 24 hours using an unpaired t-test statistical analysis. A total of 5 separate experiments were conducted with 2 treatment groups (5 untreated and 5 ultrasound-treated). One of the ultrasound treatment groups was excluded due to insufficient CO2 in the incubator as the result of a line leakage.

  The number and length of untreated and LFU-treated cells were compared at each 8-hour increment using an unpaired t-test statistical analysis. The number of HDFa cells aligned at 0˚ to 22˚ from (relatively parallel to) the wound and the number of cells oriented at a 23˚ to 45˚ angle, a 46˚ to 67˚ angle, and a 68˚ to 90˚ angle from (more perpendicular to) the wound were counted. The number of cells oriented at a 0˚ to 22˚ angle from (relatively parallel to) the wound and the number of cells oriented at a 23˚ to 90˚ angle from (more perpendicular to) the wound for both control (untreated) and LFU-treated HDFa cells at each 8-hour increment were compared using a 2-tailed Fisher’s exact test statistical analysis.

Cells treated with 35 kHz LFU exhibited a different orientation upon entry into the scratch wound compared to the untreated cells. Representative images of HDFa cell migration into the scratch wound assay at 8-hour increments for both 35 kHz LFU-treated and control cells are shown in Figure 1a. No differences in the number of cells entering the scratch wound at each 8-hour increment were observed (see Figure 2). Although the number of HDFa cells entering the scratch wounds assays did not differ between LFU-treated and untreated cells, a difference in the orientation of the cells closing the wound was observed. In most cases, a large number of the untreated cells entering the wound were aligned relatively parallel to the wound (oriented at 0˚ to 22˚ angle from the wound), while fewer LFU-treated cells were aligned relatively parallel to the wound at each 8-hour increment except for the 32-hour time point (see Figure 3).

  By the 48-hour time point, a significant difference (P <0.05, 2-tailed Fisher’s exact test) was observed between the numbers of untreated HDFa cells and 35 kHz LFU-treated HDFa cells aligned relatively parallel to the wound (see Figure 4). A greater percentage of ultrasound-treated cells (38%) were oriented 46˚ to 67˚ from the wound deficit/incision as opposed to being parallel (21%) to the wound deficit/incision. Of the remaining cells, almost equal percentages of cells in the LFU-treated group were distributed along the other orientations (21% at 0˚ to 22˚ angle, 21% at 23˚ to 45˚ angle, and 19% at the 68˚ to 90˚ angle) compared to a majority of untreated cells (46%) oriented parallel to the deficit/incision. Twice as many LFU-treated cells (34.5%) were angled at 46˚ to 67˚ compared to 17.9% for the untreated cells. Percentages of untreated cells oriented at the other angles varied widely (from 9% to 26%) (see Figure 4).

  At all time points but 1 (hour 48 post-treatment), a higher percentage of LFU-treated cells were angled from 46˚ to 67˚ (see Table 1). Similarly, at all time points but 1, untreated cells were angled predominantly at the 0˚ to 22˚ angle. Cells from both the untreated and treated groups were represented the least in the 68˚ to 90˚ angle group (14.5% and 18.4%, respectively). Similar percentages of cells from both groups were represented in the 23˚ to 45˚ angle group.

  Although a comparison of the percentages of the cultures achieving complete recoverage of the culture dish (deficit reduction/wound closure) in the untreated versus the LFU-treated HDFa at 8, 16, and 24 hours (see Figure 5) indicated no significant difference between the 2 groups, a trend toward significance was detected at hour 8 (P = 0.0543, unpaired t-test).

  Additionally, morphological observations indicated the HDFa cells exhibited different physical characteristics at the 8-hour, 16-hour, and 24-hour time points (see Figure 1a). The 35 kHz LFU-treated HDFa cells displayed a more consistently defined and elongated appearance with smaller cell bodies as compared to the control, untreated HDFa cells (see Figure 1b,c). A great amount of unoccupied space between cells was observed at the 8-hour and 16-hour time points in the untreated group. Less packing or bunching of cells in the central area of the wound deficit was observed at 24, 32, and 40 hours post-treatment, and LFU-treated HDFa cells appeared to be more evenly distributed across the wound scratch or deficit.

Fibroblasts are derived from mesenchymal tissue and are responsible for the production of collagen and other connective tissue fibers in the extracellular matrix. As such, they are the primary producer and remodeler of scar tissue after tissue injury. No previous data exist on the effects of LFU on the morphology, proliferation, and migration patterns of human fibroblast cells. However, the effects of both high-frequency and low-frequency ultrasound on other mesenchymal-derived, matrix-producing cells have been assessed. Sui-Sum Man¹⁶ examined the effects of HFU (1 MHz) and LFU (45 kHz) separately and in combination on odontoblast-like and osteoblastic cells. Both proliferation and migration rates of these connective tissue cells were studied using the wound healing or scratch assay. No change in proliferation or wound closure rates was detected for the odontoblast-like cells after 15 minutes of continuous treatment with ultrasound of any frequency. However, a significant difference in these rates was found after 15 minutes of treatment of the osteoblast cells with all frequencies (1 MHz, 45 kHz and combination of the 2 frequencies). A 5-fold increase in migration rates was detected with the combined multifrequency therapy on osteoblasts as compared to a single frequency treatment application.

  With Sui-Sum Man’s¹⁶ single-treatment frequencies of 45 kHz, migration and proliferation rates of osteoblasts were significantly higher at 24 hours compared to 48 hours. In comparison, migration rates were similar for the 24-hour and 48-hour MHz treatment group. The increased migration or wound closure rates were increased by 2.4-fold and 1.8-fold, respectively, in the MHz and KHz alone treatment groups. The effects of 35 kHz LFU on human fibroblasts were similar to the odontoblast-like cells. Neither the 35-kHz or 45-kHz LFU treatment increased the proliferation rate of these connective tissue cell types. However, in the present study, a difference was detected in the migratory pattern of the 35 kHz LFU-treated human fibroblasts (HDFa cells) at 48 hours post-treatment. Migratory patterns of the cells were not examined in the Sui-Sum Man study,¹⁶ so the effect of LFU on odontoblast-like cell migration patterns is unknown.

  As noted in the Sui-Sum Man study,¹⁶ LFU-treated cells displayed different physical characteristics: odontoblast-like cells exhibited a smaller cell body, increased spreading, and elongation when exposed to ultrasound as compared to control cells. Likewise, in this study, human fibroblasts demonstrated a more consistently elongated and defined appearance during the early time points with a seemingly smaller cell body. These physical alterations are consistent with a number of studies that indicate changes in cell morphology occur in response to mechanical stimuli.¹⁷ Work by Meazzini et al¹⁸ and Dhopatkar et al¹⁹ demonstrated mechanical strain and stress can alter morphology and size of osteoblasts and dental tissue-derived cells. It is hypothesized ultrasound may cause vibrations within the extracellular matrix that affects the actin cytoskeleton (Roovers and Assoian²⁰), thereby providing a mechanism to mediate shape change. It is also thought activation or exposure of various receptor complexes and adhesion factors on the cell surface may account for changes in cell behavior.¹²

  In the current study, the migration patterns of untreated and LFU-treated human fibroblasts were examined in a 2-D culture using the scratch or wound assay. In these conditions, fibroblasts cultured on smooth surfaces tend to “walk” across these surfaces in a random pattern.²¹ However, in the current study, a difference in migration pattern emerged — ie, a significant difference in the predominant orientation of untreated versus LFU-treated fibroblasts emerged at 48 hours post-treatment. Analysis of the percentage of cells at select angles of orientation to the wound defect/incision demonstrated the highest percentage of untreated cells was parallel to the defect, and the highest percentage of LFU-treated cells was oriented at more of an angle (46˚ to 67˚). However, as shown in Figure 3, fibroblasts from the LFU-treated group appeared to be more evenly distributed across all angles.

  These findings are interesting, given that fibroblasts produce collagen along their given path of migration. Although the findings of an in vitro study cannot be directly extrapolated to the in vivo situation, theoretically a more varied arrangement of fibroblasts should produce a reticulated or increasingly random orientation of collagen fibers. This random orientation is similar to uninjured tissue and could lead to less scarring due to decreased bundling and enhanced strength of the replacement scar tissue in wound healing.

  Fibroblasts migrate along the provisional matrix of newly injured tissue into the wound and then begin to synthesize collagen fibers in their selected location.²² These fibers replace the fibrin meshwork from the post-injury blood clot. Although the collagen produced in this process results in replacement tissue that fills the tissue defect, this newly synthesized tissue is not identical to the original tissue it replaces. Further, the architecture of the new matrix differs from that of the noninjured tissue and is typically less functional, with a strength approximately 70% of the original tissue.

  One of the most significant differences between noninjured and scar tissue is the architecture of the extracellular matrix.²³ Studies of rodent tissue reveal collagen fibers in scar tissue are present in large parallel bundles oriented perpendicular to the basement membrane.²³ Similarly, human scar tissue has been shown to contain large, parallel arrangements of collagen fibers.²⁴ However, these large collagen bundles run parallel to the basement membrane. In both rodents and humans, noninjured skin contains smaller bundles of collagen and has a reticulated or more random orientation of these collagen bundles. The degree of scarring present in healing tissues appears to be related to this increase in collagen bundle size and parallel fiber orientation.

  The results of this study suggest LFU modulates fibroblast physical characteristics and migration in culture. Given the results of this study along with the evolving evidence to support a role for LFU in facilitating chronic wound healing⁹ and increasing collagen production in wounded diabetic rats,²⁵ the potential for LFU to decrease scarring through normalizing collagen production and orientation should be investigated.

This study employed a single treatment at a single dose. Future studies are required to determine if fibroblasts would have responded differently (migrated at different rates of speed; in different patterns) to a greater dose (time length of exposure) or more frequent treatments. Fibroblasts are among the least differentiated connective tissue cell type and as a result may have responded differently to the LFU treatment.

LFU delivered at a frequency of 35 kHz produces a different pattern of fibroblast migration in vitro as compared to the control condition and alters cell morphology. The observed pattern of migration is more varied and widely distributed across multiple angles versus the predominant parallel orientation of the migrating untreated cells. Future research should focus on the possibility this LFU-mediated effect may lead to a more normalized pattern of collagen placement in the extracellular matrix due to the increased random orientation of the migrating fibroblasts.

Dr. Conner-Kerr is Dean and Professor, University of North Georgia, Dahlonega, GA. Dr. Malpass is a postdoctoral fellow, Wake Forest University Health Sciences, Winston-Salem, NC. Ms. Steele is a private therapist in Winston-Salem, NC. Dr. Howlett is a professor at Wake Forest University Health Sciences. Please address correspondence to: Dr. Teresa Conner-Kerr, Dean and Professor, 319A HNS, 82 College Circle, Dahlonega, GA 30597; email: teresa.conner-kerr@ung.edu.