Efficacy of Biophysical Energies on Healing of Diabetic Skin Wounds in Cell Studies and Animal Experimental Models: A Systematic Review

We have systematically assessed published cell studies and animal experimental reports on the efficacy of selected biophysical energies (BPEs) in the treatment of diabetic foot ulcers. These BPEs include electrical stimulation (ES), pulsed electromagnetic field (PEMF), extracorporeal shockwave (ECSW), photo energies and ultrasound (US). Databases searched included CINAHL, MEDLINE and PubMed from 1966 to 2018. Studies reviewed include animal and cell studies on treatment with BPEs compared with sham, control or other BPEs. Information regarding the objective measures of tissue healing and data was extracted. Eighty-two studies were eventually selected for the critical appraisal: five on PEMF, four each on ES and ECSW, sixty-six for photo energies, and three about US. Based on the percentage of original wound size affected by the BPEs, both PEMF and low-level laser therapy (LLL) demonstrated a significant clinical benefit compared to the control or sham treatment, whereas the effect of US did not reveal a significance. Our results indicate potential benefits of selected BPEs in diabetic wound management. However, due to the heterogeneity of the current clinical trials, comprehensive studies using well-designed trials are warranted to confirm the results.


Introduction
Thirty million children and adults in the United States have diabetes [1]. The incidence rate of diabetic foot ulcer is 6% [2], and 45% of diabetic patients die during the first year after the initial amputation [3]. Neuropathy, peripheral vascular disease and infection are the major risk factors for non-healing foot ulceration in patients with diabetes [4]. Increased inflammation and expression of matrix metalloprotiase-9, protein tyrosine phosphatase-1B in wound tissue and elevated level of serum growth factors were also found as the main factors associated with failure to heal diabetic foot ulcers [5]. Thus, treatments that manage neuropathy, ameliorate microcirculation and promote growth factor release may be helpful in treating chronic wounds or reducing their recurrence.
Biophysical energies (BPEs) are commonly used in physiotherapy daily practice [6]. BPE options for treating diabetic foot ulcers have included electrical stimulation (ES), MHz or kHz ultrasound (US), extracorporeal shockwave (ECSW), photo energies and pulsed electromagnetic field (PEMF). A systematic review reports positive findings on the use of the BPEs (ES, photo energies, and US) in managing foot ulcers [7] and peripheral neuropathy [8] in patients with diabetes. BPEs have been used to accelerate healing of chronic diabetic foot ulcers [9] and venous ulcers [10]. Moreover, BPEs may restore diabetes-associated microvascular [9] and neurological changes [11] that are important risk factors for delayed wound healing in patients with diabetes.
Despite the positive findings reported in some clinical studies, it is almost impossible to recruit homogeneous groups of patients in practice. Patients may respond differently to the same intervention due to variations in the severity of wound, location or chronicity. In contrast, the homogeneity in both experimental and control groups can be achieved in studies utilizing cell or animal models, and they also provide more insights into the mechanisms by which BPEs promote wound healing. Previous animal studies have shown that BPEs enhance macrophage migration [12] and antibacterial effects on ulcers [13]. In addition, BPEs have been shown to accelerate collagen deposition and enhance wound contraction in healthy Sprague-Dawley rats [14]. These animal model-based pre-clinical studies have brought some insights into the mechanisms of BPEs. However, it is important to note that rodent models cannot fully recapitulate human responses to BPEs due to mechanistic differences in wound healing, so findings from such studies may not be directly translated into clinical practice.
Thus far, there is a lack of updated review in the literature that evaluates the efficacy of BPEs for wound healing in cellular or animal models. The purpose of this review is to survey the current literature for studies that use cell culture and animal models to evaluate the efficacy of BPEs on diabetic wound healing, and to infer the underlying mechanisms of how BPEs promote wound healing.

Methods
This study followed the guidelines suggested by de Vries and co-worker [15] for reporting systematic reviews of animal studies.

Data Sources and Searches
The literature search for this review was restricted to published results of cellular studies and animal experiments. Databases including MEDLINE, CINAHL and PubMed were searched, covering the period from their inception to December 2018. This review was also restricted to articles published in English. Published review articles were also excluded. Keywords and Medical Subject Headings (MeSH) including PEMF, US, ECSW, ES, and LLL were combined with wound healing (limited to "cell" and "animal") (Appendix A). A manual search of bibliographic references of relevant articles and existing reviews was also conducted to identify studies not captured by the electronic database search.

Study Selection
Published studies that reported the efficacy of BPEs in treating diabetic wounds were eligible for inclusion. The inclusion criteria were as follows: • Biophysical energies • Diabetic wound • Cell or animal experiments The exclusion criteria were as follows: • Co-interventions (e.g., co-medication) • No diabetic wounds • Human studies • Systematic review or meta-analysis

Data Extraction and Quality Assessment
Literature search was conducted independently by two reviewers (RK and MC). Articles were screened according to the title, the abstract, followed by the full paper if necessary. Duplicates were checked and removed after excluding the publications that were clearly unrelated to the purpose of this study. The full text of publications satisfying the inclusion criteria was obtained for review. At all stages, whenever there were disagreements between the two reviewers, they were resolved by discussing between themselves, sometimes with a senior and experienced reviewer (GC) or the corresponding author when necessary.
Details of the studies were extracted and summarized using a data extraction sheet. Attempts were made to obtain any missing data by contacting the authors of the studies. Data from studies published in duplicate were included only once. The data collection form consisted of demographic data (author and year published), study design characteristics (experimental groups and number of animals), animal model characteristics (species, gender, and disease etiology), intervention characteristics (dosage, timing, and duration), outcomes measures and other (dropouts).

Primary Outcomes
Objective measures of healing were investigated, including the healing rate of diabetic wounds, the time for complete closure, and the proportion of subjects with wound closure within the trial period.

Search Results
Using the pre-defined keywords and MeSH, we identified 1731 publications pertaining to the use of BPEs for diabetic wound treatment in animal and cellular models. By screening the title and abstract, we obtained 135 relevant articles and retrieved the full text for 103 publications after removing 32 duplicated articles. Of the 103 articles, 21 were excluded for reasons related to the study design (n = 4), not diabetic wounds (n = 8), with co-interventions (n = 6) or human study (n = 1). Two articles were also not included due to the lack of English version [17,18]. Finally, 82 studies that specifically examined the effects of BPEs on diabetic wound healing were critically appraised. Figure 1 illustrates the trial selection process.
Significantly increased in E1 and E2 as compared to C2 (p < 0.05).

3.
Histological scores indicated ECSW-treated wounds epithelialized more rapidly and collagen fibers are more abundant at the wound site.

4.
Up-regulated significantly in E1 and E2 on Day 7 post wounding (p < 0.05).   Three-fold increase in bFGF in E as compared to C. Degree of re-epithelialization and inflammation Significant between-group difference was found in re-epithelialization and inflammation on Day 10, but not on Day 20. Histological analysis All splinted wounds were completely re-epithelized, and granulation tissue with collage fibers filled or almost filled the whole wound bed. Morphometric analysis 3.
The density of total collagen of E was significantly higher than C.

2.
Collagen I was always greater than that observed in collagen III in all groups. 3.
Significant increase in MMP-2 and MMP-9 expression in C than E. Fathabadie

Histopathological evaluation
The synthesis and organization of collagen fibers were consecutively enhanced in the 15 mW/cm 2 group. A significant difference in the number of newly formed capillaries.

Histological analysis
Significant difference in revascularization and re-epithelialization.    More bipolar and spindle-shaped fibroblasts in the laser-treated cultures than in the sham-exposed.

2.
Significant increase in the number of cells in E5.

3.
Ultrastructure features of fibroblasts in the sham-exposed and laser-treated cultures were similar. Increase in cellular viability.

4.
All cells model showed an increase in proliferation.

5.
Decrease in TNF-α and proinflammatory cytokine interleukin IL-1β. E3 showed a decrease in TNF-α. Increase in S-phase and decrease in G2M phase. TGF-β1 significantly more elevated after irradiation than sham-irradiated controls.

2.
Greater increase in MMP-2 was found after irradiation than sham-irradiated controls. MTS assay 3.

Polychromatic light emitting diode (LED) energy
CyQuant assay 1. Increase in total adenosine triphosphate production at both power densities except the power density of 10 mW/cm 2 and 5 J/cm 2 .

2.
Mitochondrial metabolism was significantly higher.

3.
Significantly higher cellular proliferation with groups irradiated with 10 mW/cm 2 . E, Experimental group; C, Control group.
Wound closure percentage was the main outcome measure for all five trials. Other measures included overall wound closure time, cell proliferation, vascularity, murine endothelial cell culture, FGF-2 secretion, wound tensile strength, myofibroblast production, type 1 collagen fiber deposition, collagen fibril alignment, collagen fiber anisotropy and orientation, energy absorption capacity, Young's modulus, wound thickness, and maximum stress of wound tissue. Four trials report significant between-group difference in the percentage of original wound size, and the experimental groups in all these studies demonstrated improved wound healing compared to the control groups [19][20][21]23].

Ultrasound (US)
Two trials compared ultrasound with sham treatment [24,26], whereas one trial compared ultrasound with dressing changing [25]. The wound size was the main outcome measure for all three ultrasound trials. Other measures included wound closure duration, granulation tissue, collagen deposition, angiogenesis, VEGF expression, SDF-1 expression, fibroblast proliferation, speed and persistency of fibroblast migration ( Table 2).
Male CD-1 mice, BKS.Cg-Dock7m+/+Leprdb/J mice, Syndecan-4 wild-type and knockout C57BL/6J mice were used in the animal models. Fibroblasts from wound tissues and db/db mouse skins were used as the cellular model. Thawer et al. and Mann et al. delivered ultrasound with saline vapor at 45 kHz and 40 kHz, respectively, while Roper et al. delivered 1 kHz ultrasound through water-based gel. Two out of three trials revealed significant between-group differences in wound size in favor of the experimental groups over the control groups in these studies [25,26]. The exception was the trial reported by Thawer and collaborators, which showed no significant between-group differences in wound size after ultrasound treatment.
Kuo and colleagues compared three different protocols of shockwave with the control group receiving no shockwave energy and reported a significant acceleration in wound healing (p < 0.05). The perfusion in wound area was significantly higher in the experimental group treated with two sessions of defocused shockwave (on postoperative Days 3 and 7) than the diabetic control group (p = 0.023). In addition, fibroblast count and VEGF level were upregulated in experimental groups compared to control groups. The authors concluded that treatment with an optimal session of ECSW significantly enhanced diabetic wound healing associated with increased neo-angiogenesis, tissue regeneration and topical anti-inflammatory response. However, they did not provide details on the randomization method, allocation concealment, random housing, outcome assessment, and investigator and assessor blinding [27]. Yang and colleagues compared two different protocols of shockwave with the control groups, and they reported a significant improvement evident by increased wound breaking strength, number of fibroblasts and collagen fibers. The authors concluded that low energy ECSW can improve the healing of incisional wound in diabetic rats [28]. Zins et al. investigated the angiogenic gene expressions and wound closure kinetics during diabetic wound healing with or without ECSW therapy. The expression of certain genes in the diabetic wound was augmented by shockwave, especially PECAM-1; however, they found that shockwave had no effect on wound closure in both normal and diabetic models [30].
Hayashi et al. investigated the role of endothelial nitric oxide synthase with shockwave energy for diabetic wounds. A single session of ECSW accelerated wound healing in a streptozotocin-induced diabetic mouse model, accompanied by an increased expression of eNOS and vascular endothelial growth factor (VEGF). However, the efficacy of ECSW was attenuated in eNOS-KO mice. The authors concluded that eNOS played a critical role in the therapeutic effects of shockwave by accelerating the wound healing through VEGF upregulation and neovascularization [29].

Electrical Stimulation (ES)
The four ES trials used different types of protocols. Two trials compared ES with sham treatment [32,33]. One trial compared two different ES protocols with control receiving no ES [31]. Another trial compared ES with the control group receiving no ES or with transdermal iontophoresis by zinc sulfate [34]. None of these studies provided information about randomization, allocation concealment, investigator and assessor blinding, random housing and outcome assessment (Table 4).
Monophasic pulse wave is reported in two trials [32,33]. The outcome measures included wound healing rate, wound contraction, tensile strength, histology, collagen deposition, fibroblast proliferation and morphological analysis. Smith et al. classified the tensile strength into "poor", "moderate" and "good" after 10 days of stimulation, and they showed that ES enhanced diabetic wound healing. However, no statistical analysis is provided in their study [31].
Thawer et al. compared wound healing in diabetic mice with ES at 12.5 V and sham treatment (0 V). No statistical difference was found in epidermis thickness between groups. The authors suggested that ES at a high dose can alter collagen deposition in excisional wounds of diabetic mice; however, they found the effect of ES on wound healing to be disease-specific [32]. Kim and colleagues compared experimental groups receiving ES at 35-50 V with a control group receiving sham ES. Significant difference was found in wound healing rate between groups. In addition, elevated levels of collagen-I, α-SMA and TGF-β1 were found in experimental groups (all p < 0.05) [33].
Langoni Cassettari and collaborators divided the normal and diabetic Wistar rats into six experimental groups to study the effect of ES with direct current (DC) and zinc sulfate treatment by transdermal iontophoresis. The authors concluded that DC alone or used in association with zinc by transdermal iontophoresis was able to induce the morphological and ultrastructural changes observed during surgical wound healing in diabetic animals [34].

Low Level Laser Therapy (LLL)
A broad spectrum of laser wavelengths has been reported by different studies, whereas wavelengths in the visible red range (630-685 nm) were most commonly investigated either in isolation or in combination with other wavelengths ranging from 425 nm to 1064 nm. Power density in mWcm 2 was not specified in some of the reviewed studies, even though this represents an important parameter. The irradiance ranged widely from 4 to 79 mWcm 2 . Peplow et al. reported a range of irradiance instead of a specific density [71]. Similarly, a large variety of animal models have been used, including C57BL/Ksj/db/db mice, SD rats, Sand rats, Wistar rats, BKS.Cg-m+/+Leprdb/J mice, Zucker diabetic rats and Swiss albino mice. Several wound healing outcomes were measured using various techniques, most commonly wound size and histology. However, nine of our surveyed trials applied laser to parts of the body of a single animal for both experiment and control, and analysis was conducted as if every single wound were from an individual animal [37,40,42,59,60,74,78,82,83,89].

Polychromatic Light Emitting Diodes (LED)
In six trials that investigated effects of polychromatic light emitting diodes (LED), three trials studied burn healing in diabetic rats [39,67,70]. Al-Watban et al. compared the efficacy of LED (wavelength 510-543, 594-599, 626-639, 640-670 and 842-879 nm) on burn wound at four different doses with the sham treatment. Significant burn healing was found from 48.77% to 76.77% after LED stimulation at different doses in diabetic rats [39]. The same research group also compared the efficacy of laser of different wavelengths (532, 633, 810, 980, 10600 nm) to LED clusters (510-872 nm) with incident doses of 5, 10, 20 and 30 J/cm 2 in SD rats (n = 893) [56]. Their results showed that phototherapy at 633 nm should be given three times a week at a fluence of 2.35 J/cm 2 each time for diabetic wound treatment. Wu et al. [91] compared the 635 nm laser with organic LED and showed that the organic LED significantly increased fibroblast growth factor-2 expression and macrophage activation during the initial stages of wound healing. In addition, they also found that organic LED and laser had comparative effects on promoting diabetic wound healing in rats.

Infrared (IR)
Danno and colleagues conducted both in vitro and in vivo studies to compare the infrared irradiation treatment with sham irradiation control or thermal control [36]. The TGF-β1 and MMP-2 content in the medium of cultured cells was significantly elevated after irradiation. Negative results in thermal controls suggested that the action of the light was athermic in nature. In animal models, the rate of wound closure was significantly accelerated after repeated exposures. Cheing and collaborators compared the efficacy of managing acute wounds in male diabetic SD rats between groups of monochromatic infrared energy (MIRE) at 890 nm and the sham group without receiving infrared energy [79]. Both experimental and sham groups showed improvement in terms of wound closure percentage; however, no statistical difference was found between groups.

Discussion
Preclinical research is important for expanding knowledge and provides insights into the cellular and physiological mechanisms on how BPEs enhance diabetic wound healing. Two trials have investigated how cells respond when exposed to electrical currents [101,102]; however, research evidence showing its effects on diabetic wound healing is limited. Four in vivo studies described here present inconsistent results regarding the value of ES in acute diabetic wound healing in animals. Thawer et al. showed no statistical difference in epidermis thickness between groups, but they did find a significant increase in collage deposition [32]. Findings reported by Kim et al. are consistent with those found by Thawer's team, in which collagen-I expression was higher after ES. In addition, α-SMA and TGF-β1 expression were also enhanced after daily ES [33]. Langoni Cassettari et al. found accelerated wound contraction, but the morphology of inflammation was not altered after ES [34]. Statistical analysis was not available in one of the studies examined [31], making it difficult to draw conclusions on the ES' benefits in diabetic wound healing from this animal study.
Extracorporeal shockwave (ECSW) has been used clinically for treating musculoskeletal disorders and diabetic ulcers for some years [103]. However, preclinical studies examined in this review reported contradictory findings in supporting the use of ECSW on diabetic wound healing. Two studies showed that ECSW significantly reduced wound size compared to sham treatment groups in diabetic rats [27,29]. On the contrary, a recent study by Zins et al. found that ECSW did not accelerate wound closure in wildtype (nob-diabetic) mice or db/db diabetic mice [30]. Another study found that diabetic mice treated with ECSW significantly increased the wound breaking strength and the collagen fiber content [28]. However, this effect was attenuated in endothelial nitric oxide synthase-knockout mice, suggesting that nitric oxide synthesis plays a critical role in the therapeutic effects of ECSW in diabetic wound healing [29].
Pulsed electromagnetic field (PEMF) energy has been used to treat diabetic stump wounds [104] and chronic diabetic ulcers [9]. All five studies included in our review showed positive findings that supported the use of PEMF in promotion of diabetic wound healing in animal models [19][20][21][22][23]. However, when Callaghan et al. repeated the same protocol on FGF-2 knockout mice, there was no significant improvement found in wound closure rate, suggesting FGF-2 might be a crucial factor in PEMF stimulated diabetic wound healing [19].
Sixty-six studies concerning photo energies are included in the present review. Different types of photo energies with different frequencies have been used in various studies. The wavelengths used range from visible red to infrared, power values from milliwatt to watt, and irradiation from seconds to hours. The wide range of irradiation parameters from the current review suggests the bio-modulatory potential of laser therapy [105]. In addition, these studies were conducted using various diabetic wound models, and different outcome measures were used. The findings show that irradiation by laser accelerated wound closure and collagen production, and there were increases in cellular migration, tissue viability, growth factors and gene expression. Histopathological analysis also showed a decrease in inflammatory cells and an increase in vascularization after irradiation compared to the sham control. Most trials report positive results, except Jahangiri Noudeh et al. who found no statistical significance by repeated measurements throughout the entire study period when a combined 670 nm and 810 nm laser was applied to wound areas [66]. Histological analysis revealed that there was an increase in macrophages [61,95,99], fibroblasts [47,53,63,67,68,81,84,99,100], neutrophils [95], T lymphocyte [95], collagen deposition [37,40,70,77,82,85,99,100], nitrite [100] and nitric oxide level [65], catalase activity [100], thiobarbituric acid reactive substances [100] and vascularization [44,68,70,99] after irradiation. Chung et al. adopted a splinted diabetic wound model to minimize mouse skin contraction during wound healing [62]. Seven-day treatment of 3.7-5.0 J/cm 2 caused maximum stimulation of wound healing in diabetic mice compared to the mice receiving no irradiation. Laser irradiation of wavelength at 780 nm improved muscle repair by enhancing reorganization of myofibers and perimysium in cryoinjured diabetic rats [87]. However, not all studies demonstrated a positive result due to the specificity of absorption spectrum and laser intensity. For instance, higher frequencies might cause a negative effect on cells. Houreld and Abrahamse compared the cell morphology and expression of human IL-6 between groups receiving 5 and 16 J/cm 2 . They found that subjects treated with 16 J/cm 2 demonstrated signs of stress without a significant increase in IL-6 expression [51]. Therefore, the optimal protocol of laser therapy for enhancing diabetic wound healing should be further investigated.
The present review does not support the use of ultrasound (US) in promoting diabetic wound healing using animal models [24][25][26]. Thawer and collaborators [24] demonstrated no significant between-group difference in wound size reduction after US, however, a significant improvement was shown by Mann et al. and Roper et al. after treatment [25,26]. Fibroblast migration and proliferation [24][25][26], as well as vascular density [24,25], were enhanced by the use of US compared to the sham groups. Interestingly, these two studies applied 40 and 45 kHz US to wounds through saline vap or or mist (as the coupling medium) for 1.5 and 3 min, respectively [24,25]. Another study utilized US at 1.5 MHz applied via traditional coupling gel for 20 min [26]. The optimal protocol for using ultrasound for enhancing diabetic wound healing should be further evaluated in future studies.
Most research on BPEs have been conducted on animal models consisting of surgically excised skin or burn wounds. However, no animal tissue model could possibly replicate the clinical situation in humans because different species may involve different healing mechanisms in skin wound, therefore, treatments with different BPEs are likely to yield different cellular responses when compared to human skin [106]. These experimental wounds excluded common problems associated with delays in healing including ischemia and infection, thus they might not present the real situation in humans [107]. In addition, Wang et al. commented that most in vitro data derived from fibroblasts of abnormal wound lesions only represent the terminal stage of the disease [107]. Therefore, these wound models may not be ideal to study the effect of BPEs on human diabetic ulcer healing. Recently, a reproducible chronic diabetic wound model that had low mortality rate was established by using Pseudomonas aeruginosa biofilm in db/db mice [108,109]. This model could be adopted in future studies to evaluate the antibiofilm effectiveness of BPEs in chronic wounds, which simulate infected diabetic ulcerations commonly seen in clinical settings. It should be noted that humane issue is always a concern of animal studies, in particular for experiments involving burn and wound. Therefore, in vitro methods might be an alternative because not only the humane concerns are circumvented but also the human cells instead of animal cells can be directly tested. Due to the shortcomings of animal studies, well-designed human studies are still the gold standard in clinical practice.

Conclusions
The present review demonstrates methodological shortcomings in animal studies that have studied the efficacy of BPEs in diabetic wound healing. One major limitation exhibited in animal experiments is that random allocation of animals to experimental and control groups and blinding is not yet a standard practice [110]. In addition, critical information for animal housing conditions and dropouts are unreported. Investigators should consider the findings of this systematic review when designing future studies and attempting to improve the internal validity of the studies by using true randomization in group allocation and outcome assessment, investigator and assessor blinding, allocation concealment, random housing, and reporting accurately on the number of animals used. In this review, the search was restricted to English publications as the translation was not available for full text review, which may have resulted in language bias. Notably, a variety of animal models were used for in vivo wound healing studies, but the physiology and healing mechanisms may not be the same across different species, and they are even more distinct compared to humans. There was considerable variation in research design, methodology, and parameters which limited comparison of research findings between studies. Therefore, findings obtained from even well-controlled animal studies may not be readily translated into clinical practice for people with diabetes management. Based on positive effects of PEMF and photo energies towards diabetic wound healing, more high-quality human clinical trials to assess the effects of those biophysical energies are warranted in the future.

Conflicts of Interest:
The authors declare no conflict of interest.