Biological effects of a new ultraviolet A1 prototype based on light‐emitting diodes on the treatment of localized scleroderma

Ultraviolet A1 (UVA1) phototherapy (spectral range 340‐400 nm) is a well‐established treatment option for various skin diseases such as localized scleroderma. Recent improvements of conventional UVA1 light sources (metal‐halide or fluorescent lamps) have brought attention to a new light‐emitting diode (LED) technology with remarkable advantages in handling and clinical routine. This study provides a preclinical histological and molecular evaluation of an LED‐based UVA1 prototype with a narrower spectral range (360‐400 nm) for treating localized scleroderma. Scleroderma mouse models and fibroblasts in vitro were exposed to LED‐based UVA1 phototherapy or to irradiation with a commercially available metal‐halide lamp emitting low‐dose (20, 40 J/cm2), medium‐dose (60 J/cm2) and high‐dose (80, 100 J/cm2) UVA1 light. Both UVA1 light sources affected inflammatory genes (IL‐1α and IL‐6) and growth factors (TGFß‐1 and TGFß‐2). Increased collagen type 1 was reduced after UVA1 phototherapy. Matrix metalloproteinase‐1 was more enhanced after a medium dose of LED‐based UVA1 phototherapy than after conventional treatment. In vivo, dermal thickness and the amount of collagen were reduced after both treatment methods. Remarkably, myofibroblasts were more effectively reduced by a medium dose of LED‐based UVA1 phototherapy. The study indicates that LED‐based UVA1 phototherapy yields similar or even better results than conventional treatment. In terms of biosafety and patient comfort, LED‐based UVA1 phototherapy offers clear advantages over conventional treatment because of the use of a narrower and less harmful UVA1 spectrum, less heat generation and shorter treatment times at the same irradiation intensity. Clinical studies are required to confirm these results in patients with localized scleroderma.


| INTRODUC TI ON
Localized scleroderma (LS) is a distinctive inflammatory disease that leads to sclerosis of the skin and subcutaneous tissues. [1][2][3][4] Management of LS is difficult, and no causal therapy is available so far. Topical and systemic immunosuppressants such as corticosteroids are commonly used as anti-fibrotic agents but with considerable side effects and limited efficacy. [5,6] The beneficial effects of different forms of phototherapy in the treatment of LS have been described in many studies. [7][8][9][10][11][12][13] The development of UVA 1 phototherapy (340-400 nm) has highly improved treatment of LS. [14,15] Because UVA 1 light is less erythematogenic than broadband UVA, it is possible to apply much higher doses of UVA 1 without the risk of sunburn.UVA 1 light penetrates deeper into the skin than UVB and UVA 2.
[ 16] In the first prospective study on LS, high-dose UVA 1 light (130 J/cm 2 ) was highly effective in reducing skin sclerosis, and some patients even showed complete clearance. [11] Lower UVA 1 doses (30 J/cm 2 ) were also highly effective and well-tolerated by patients with LS who showed a significant reduction in skin thickness. [17] Commercially available UVA 1 light sources for therapeutic use are fluorescent lamps (ie TL10R 100W, Philips) or high-output metal-halide lamps (ie Sellamed 4000W, Sellas Medizinische Geräte GmbH). Lowdose to medium-dose UVA 1 phototherapy is usually conducted by means of fluorescent lamp cubicles, that is full-body UV cabins, allowing the irradiation of large body areas; in contrast, metal-halide lamps are more often used for high-dose UVA 1 treatments. [14,18,19] Despite to be treated. The narrower UVA 1 spectrum due to the mixture of different LEDs is also of importance. Patient safety could be improved by eliminating potentially more damaging short-wave UVA 1 radiation (340-360 nm). This study investigated the-so far-unresolved question whether molecular or cellular mechanisms are affected by the different UV spectrum.

| Cell culture
The in vitro study included human dermal fibroblasts from n = 6 dif-

| Animals and treatment groups
The mouse experiments were conducted with129Sv/Ev male mice  (Table S1). To induce a comparable level of local fibrosis that was as homogeneous as possible, the injections were always given by the same person. Groups 3-6 received phototherapeutic treatment for 30 days (5 times per week). The high-dose UVA 1 groups (groups 3 and 4) were treated for 3330 minutes (= 100 J/cm 2 ) per session and the medium-dose UVA 1 groups (groups 5 and 6) for 2000 minutes (= 60 J/cm 2 ). Groups 3 and 5 were exposed to irradiation with the metal-halide lamp and groups 4 and 6 to irradiation with the LED prototype (Table S1). The mice were treated in their usual cages to avoid any additional stress. After UVA 1 therapy, all animals were killed by cervical dislocation, and the bleomycin-treated or PBS-treated skin area was removed and subdivided for histological preparation or molecular biological examination.

| UVA 1 devices, illumination set-up, and treatment of cells and mice
This UVA 1 phototherapy study included two UVA 1 devices ( Figure   S1A,B) that differed with regard to the type of lamp and the UVA 1 spectrum ( Figure S1C). Besides the conventionally used UVA 1 device

| Quantitative real-time PCR analysis
Cellular gene expression analysis consisted of quantitative realtime PCR with specific sets of primers (Sigma-Aldrich) and conditions (Table S2) using LightCycler technology (Roche Diagnostics) as described elsewhere. [24] PCR reactions were evaluated by melting curve analysis. ß-actin was amplified to ensure cDNA integrity and to normalize expression. Each real-time PCR was repeated at least 3 times.

| Enzyme-linked immunosorbent assay (ELISA)
Cell supernatants were collected 48 hours after UVA 1 treatment and of untreated control fibroblasts and were analysed by total MMP-1

| Histochemical analysis
Immunohistochemistry was conducted on formalin-fixed and paraffin-embedded full-skin preparations from all mice (n = 6 mice per group H&E staining on murine skin sections was done as described previously. [26] Extracellular matrix accumulation in skin sections was determined by staining with Sirius red/fast green. [27] The intensity of the Sirius red staining of the murine tissue sections was quantified with the image J software (http://www.imagej.softo nic.de) and expressed as a percentage of the total area. All tissue slides were scanned with a histo-scanner (PreciPoint, M8), and representative pictures were visualized.

| Analysis of collagen content
Collagen in mouse tissues was analysed with the QuickZyme collagen assay according to the manufacturer's instructions (QuickZyme Biosciences).

| Statistical analysis
Data are expressed as mean values ± standard deviation (SD). Data were analysed with GraphPad Prism 5 software (GraphPad Software Inc). Groups were compared by one-way ANOVA with Tukey's honestly significant difference (HSD) post hoc test analysis. Differences were considered statistically significant at P < .05.

| Study approval
The animal study was conducted according to the NIH Guide for the

| RE SULTS
The study purpose was to evaluate molecular and histological ef-  Figure S1 shows the exposure unit and the emission spectrum of the UVA 1 metal-halide high-pressure lamp (SELLASOL) ( Figure S1A, C) and the LED-based UVA 1 prototype ( Figure S1B,C).

| Examination of morphology and cell vitality of human fibroblasts after UVA 1 treatment
To compare the two UVA 1 light sources with respect to cell viability, cell morphology and cell vitality were determined 24 hours after UVA 1 treatment by means of an MTT test. Human fibroblasts exposed to high UVA 1 doses (80 and 100 J/cm 2 ) appeared swollen, broken and round, and the damage was most critical after treatment with either light source ( Figure 1A). Single UVA 1 treatments with up to 60 J/cm 2 did not change the morphology of the fibroblasts.
Cell vitality significantly decreased with increasing UVA 1 doses ( Figure 1B). Doses > 60 J/cm 2 were defined as the irradiation limit for both UVA 1 light sources and were not used in further cell culture experiments.

| Effects of UVA 1 on cytokines, growth factors and collagen metabolism in fibroblasts
mRNA expression of IL-1α and IL-6, which was initially analysed 6 hours after exposure to UVA 1, was induced after treatment with both the metal-halide lamp and the LED-based prototype in comparison with untreated control. mRNA expression did not seem to be dependent on the UVA 1 dose (Figure 2A The collagen content in the affected skin after UVA 1 treatment was histologically analysed by means of Sirius red/fast green staining ( Figure 4B) and image analysis ( Figure 3B) and quantified using a hydroxyproline assay ( Figure 3C). The amount of collagen was reduced after each UVA 1 treatment regimen; however, compared to the bleomycin-only group (group 2), the reduction was only significant after high-dose treatment with either light source ( Figure 3C). Finally, the number of α-SMA-positive myofibroblasts was determined by a dermato-histopathologist. Myofibroblasts were significantly reduced after high-dose UVA 1 treatment with the metal-halide lamp and medium-dose UVA 1 treatment with the LED-based UVA 1 prototype ( Figure 3D). The obviously stronger α-SMA staining in groups 3-6 ( Figure 4C) was due to the blood vessels that were also α-SMA positively stained and whose synthesis is known to increase after UVA 1 irradiation . [28] In summary, these in vivo results indicate that both UVA 1 light sources are effective F I G U R E 2 Gene expression marker analyses in human fibroblasts after UVA 1 treatment. Different fibroblasts (n = 6 different donors) were exposed to UVA 1 (metal-halide UVA 1 or LED-based UVA 1 ) using different treatment regimens (20,40 and 60 J/cm 2 ) or remain untreated (ctrl.; 0 J/cm 2 ). RNA was isolated 6 or 24 h after treatment as indicated in the text, and different sclerosis-related molecular markers were analysed. A, IL-1α, (B) IL-6, (C) MMP-1, (D) Col-1, (E) TGFß-1 and (F) TGFß-2 (* # P < .05, ANOVA with Tukey's multiple comparison test) in reducing sclerosis in localized scleroderma with minor differences according to the dose of UVA 1 irradiation applied.

| D ISCUSS I ON
The aim of this study was to evaluate the impact of a new LED- mice, [31] but the optimum UVA 1 spectrum for scleroderma treatment has not yet been determined and needs to be examined further.
To determine possible differences between metal-halide lamps and the LED-based UVA 1 prototype on the molecular level, we ana-  [37,38] Part of these molecules efficiently generate singlet oxygen which in turn may induce MMP-1. [39,40] This hypothesis is purely speculative and must be investigated in further studies.
In addition to MMP-1 induction, our study also showed reduction in Col-1. Although collagen balance is of particular importance in sclerotic diseases, collagen degradation by UVA 1 treatment plays a relatively minor but not less important role than the induction of MMP-1. In percentage terms, UVA 1 irradiation has a significantly stronger effect on MMP-1 activation than on the reduction in Col-1 [32] as confirmed by our study.
In addition to the intervention in the collagen balance of fibroblasts, another important mechanism of UVA 1 irradiation is its immunomodulatory effect. This effect particularly concerns pro-inflammatory cytokines such as IL-1α and IL-6 that signal mediate functions inside and outside a cell. Stimuli of cytokine production in the cells are different stress factors such as exposure to UV light. The synthesis of pro-inflammatory cytokines forms a mutually stimulating network, in which IL-1α plays a special role.
IL-1α may influence the collagen metabolism and is a major stimulus for the synthesis of other profibrotic regulatory proteins such as IL-6 or IL-8. [41,42] Kreuter and colleagues described significantly down-regulated mRNA expression of IL-6 and IL-8 after UVA 1 treatment in patients with LS. [43] Vielhaber et al controversially discussed in their study that increased UVA 1 -induced MMP-1 production is connected with the upregulation of the cytokine IL-1α and the resulting increase in IL-6 synthesis. [44] Increased IL-1α and IL-6 gene expression after UVA 1 irradiation of fibroblasts was also found by Wlaschek et al [36,40] Any comparison of study results should distinguish between normal fibroblasts and LS-derived fibroblasts. Our study involved normal fibroblasts. Similar to the results of Wlaschek et al, [36,40] we observed a significant induction of IL-1α and IL-6 expression for both UVA 1 light sources. In addition, the growth factors of the TGFß family are considered to be further important indicators for fibrosis and play an essential role in scleroderma. In the dermis, TGFß stimulates the proliferation of dermal fibroblasts, which in turn secrete increased amounts of ECM components such as Col-1 and have a reducing effect on MMP-1 expression. [45][46][47][48] In addition, TGFß may suppress the production of IL-1, TNFα and IL-8, which in turn shows the interplay between pro-and anti-inflammatory functions in cytokine-mediated inflammatory reactions. [49] The two most important isoforms of TGFß are TGFß-1 and TGFß-2, both crucial factors in the pathogenesis of fibrosis. [47,[50][51][52][53] In activated scleroderma fibroblasts, the growth factor TGFß-1 is excessively secreted. [45] In addition, mRNA gene expression of TGFß in lesional skin areas significantly decreases after UVA 1 therapy in patients with LS. [45] Other in vivo studies using normal human skin samples yielded contradictory results. Quan et al described the significant upregulation of TGFß-1 combined with a simultaneous decrease in TGFß-2 after exposure to UV irradiation. [54] In experiments with skin fibroblasts, Yin et al also observed the significant induction of TGFß-1 after UVA treatment. [48] The results of Quan et al and Yin et al [48,54] are also reflected in our study. 24 hours after UVA 1 exposure, TGFß-1 expression was increased simultaneously with reduced TGFß-2 expression in normal fibroblasts.
Effective UVA 1 treatment of scleroderma in vivo decreases dermal thickness primarily due to the degradation and reduction in collagen. [32] In our in vivo experiment, all groups except control group 1 received bleomycin injections and developed the typical picture of the fibrosis mouse model described by Yamamoto et al [20][21][22][23] The bleomycin control group 2 clearly showed thicker dermis than groups 3, 4, 5 and 6 that were additionally treated with UVA 1 . Also clearly visible were the fat cells displaced by dermal fibrosis that resulted in complete absence of these cells due to highly pronounced fibrosis, especially in control group 2 of this study. Here, fat cells are replaced by fibrotic tissue through adipocyte-myofibroblast transformation. [55,56] Slight regeneration of fatty tissue was observed in the experimental groups treated with UVA 1 . Increased myofibroblast activity and associated profibrotic processes significantly contributing to the development of fibrosis [21,57] could be observed in the α-SMA stains of Karpec and colleagues described in their animal study that even a cumulative dose of UVA 1 with a narrowband (365 ± 5 nm) LEDbased UVA 1 light source that was 5 times lower than that of conventional broadband UVA 1 lamps had almost the same effect on decreasing dermal thickness (skin thickness was reduced by 34% after irradiation with a cumulative dose of 600 J/cm 2 and by 36% using 3000 J/cm 2 ). [31] In our similar mouse model, reduction in the dermal diameter in the 129Sv/Ev mouse strain was lower than that achieved by Karpec using two different devices for UVA 1 irradiation. This difference may be due to the different animal strains Further clinical studies are required to confirm these results in patients with LS or related sclerotic skin diseases such as eosinophilic fasciitis, lichen sclerosus et atrophicus or chronic sclerodermiform graft-versus-host (GvHD) disease, in which UVA 1 phototherapy is also a treatment option.

ACK N OWLED G EM ENTS
The authors acknowledge Dr Robert Sellmeier and Sellas Medizinische Geräte GmbH for providing the UVA 1 devices. We thank Monika Schöll for the linguistic revision of the manuscript.

CO N FLI C T O F I NTE R E S T
The authors have no conflict of interest to declare.