Article ID: 2025-011
Narrowband ultraviolet radiation (NB-UVR) phototherapy is widely used to treat various skin diseases such as atopic dermatitis and psoriasis. Although the mechanisms of NB-UVR at doses exceeding the minimal erythema dose (MED) have been extensively studied, the mechanisms of sub-MED irradiation remain unclear. This study investigated the effects of NB-UVR on naturally occurring canine atopic dermatitis and analyzed the underlying mechanisms of sub-MED irradiation using in vivo and in vitro canine models. Clinically, NB-UVR irradiation at 310 and 320 nm improved lesion scores, with a significant effect observed with low-dose 320 nm irradiation (n=6, P<0.05). To compare the effects of sub-MED of NB-UVR wavelengths (310, 320, and 330 nm), intradermal skin tests in experimental beagles showed significant alleviation of anaphylactoid reactions across all wavelengths (n=3, P<0.01). The number of cutaneous mast cells was significantly reduced following 310 nm irradiation (n=3, P<0.05). In vitro, WST assays and DAPI staining demonstrated reduced cell viability at 310 and 320 nm (P<0.01), with apoptosis specifically induced at 310 nm (P<0.01). RNA sequencing identified differentially expressed genes associated with muscle cell components and function after 320 nm exposure (n=3, q<0.1). These findings suggest that sub-MED NB-UVR, especially at 320 nm, may offer non-erythemogenic and safer phototherapy, with potential benefits for skin repair.
The immunomodulatory effects of sub-MED narrowband ultraviolet radiation (NB-UVR) are less well understood than those of NB-UVR above MED levels. This study shows that 320 nm low-dose NB-UVR improved clinical signs, especially hair regrowth, in dogs with naturally occurring atopic dermatitis, a valuable human disease model. In canine in vitro and in vivo models, distinct 310 and 320 nm wavelengths of sub-MED NB-UVR had specific therapeutic mechanisms: both reduced skin anaphylactoid reactions, but 310 nm induced apoptosis, whereas 320 nm promoted muscle-related gene expression. Our findings suggest that 320 nm NB-UVR represents a non-erythemogenic phototherapy approach, complementing the known apoptotic mechanisms of NB-UVR.
Ultraviolet radiation (UVR) is widely utilized in therapeutic applications for skin diseases because of its immunomodulatory action. The wavelengths used in skin phototherapy include ultraviolet (UV) A2 (315–340 nm), UVA1 (340–380 nm), broadband UVB (280–315 nm), and narrowband UVB (308–313 nm). These treatments have demonstrated efficacy in managing conditions such as psoriasis, atopic dermatitis (AD), urticaria, mastocytosis, mycosis fungoides, and vitiligo [1,2,3]. Efforts to enhance the safety and efficacy of UVR therapy have focused on identifying precise UV wavelengths tailored to specific diseases and developing narrower-bandwidth illuminators [2, 3]. Recent advancements, including excimer lamp (308 nm) and light-emitting diode (LED) technologies, have refined targeted phototherapy applications [3]. However, the specific therapeutic effects of narrower wavelengths, such as those separated by 10-nm intervals, remain poorly understood.
The mechanisms underlying the effects of therapeutic UVR doses have been extensively studied, particularly those of the keratinocyte damage pathway, which plays a key role in immunomodulation. UVR is absorbed by chromophores, leading to the activation of keratinocytes and immune cells such as the Langerhans cells and mast cells (MCs) [1, 3]. The activated cells produce inflammatory (TNF), Th2 (IL-4), and anti-inflammatory (IL-10) cytokines; reduce antigen-presenting cell function; alter regulatory T cell populations; and suppress systemic contact hypersensitivity (CHS) [1,2,3]. Since the early 2000s, several studies have investigated UV-induced immunosuppression in MCs and their mediators, using models such as bone marrow-derived MC-transplanted and MC-deficient mice [4,5,6,7]. However, despite the importance of these pathways in the overall immune response, most studies have predominantly focused on UVR doses exceeding the minimal erythema dose (MED) [1, 8, 9]; however, the significance of sub-MED doses remains poorly understood.
In basic research on narrowband ultraviolet radiation (NB-UVR), mice have frequently been used; however, an ongoing debate exists regarding whether therapeutic wavelengths optimized for humans are equally suitable for mice, given the differences in skin structures and ex vivo studies [10] reporting potential confounding factors, such as altered blood flow and surgical tissue damage. In contrast, canine models are valuable natural models for studying AD [11,12,13] and offer several advantages over rodent models. Canine MCs, which play a central role in canine AD, are biologically similar to their human counterparts in terms of origin and serine protease expression, including tryptase and chymase [11, 14]. Canine AD is one of the most common diseases in small animal veterinary dermatology [15]. Among the existing treatment guidelines [15], the mechanisms of NB-UVR therapy, established in humans, have been explored in dogs as a novel therapeutic approach [16, 17], and its efficacy has recently been demonstrated in cases of canine AD [18]. Research on NB-UVR in dogs not only contributes to the treatment of canine AD but also provides a valuable preclinical model for human medicine.
In the present study, we evaluated the effects of sub-MED NB-UVR in dogs. First, we assessed the clinical efficacy of NB-UVR using a novel LED device in naturally occurring canine AD. Next, we investigated the effects and underlying mechanisms of sub-MED NB-UVR at different wavelengths (310, 320, and 330 nm) using experimental beagles. Additionally, we examined cell viability and apoptosis induction under each irradiation condition using a MC line.
This study was approved by the Institutional Animal Care and Use Committee of Azabu University (approval number: 180924-5) and conducted in compliance with the ARRIVE guidelines 2.0. Twelve dogs were included in this study: six client-owned dogs for the clinical trial (Supplementary Table 1) and six experimental beagles for the basic experiments (Supplementary Table 2).
The six client-owned dogs with spontaneous AD were enrolled in the NB-UVR clinical trial (Supplementary Table 1). Canine AD was diagnosed based on Favrot’s criteria [19]. Informed consent was obtained from all owners before the trial commenced.
The remaining six dogs were healthy beagles (two males, four females; mean age: 5.8 years) obtained from Kitayama Laboratories (Nagano, Japan). These dogs were first used to determine the MED (n=6). Subsequently, the effects of repeated sub-MED NB-UVR irradiation were evaluated using an intradermal skin test (IDT) (n=3), histopathological analysis (n=3), and the whole transcriptome RNA sequencing (RNA-seq) of the skin (n=3). The beagles were housed in the institute’s laboratory animal facility under appropriate conditions to ensure a safe, clean, and spacious environment, with proper feeding and temperature control. Routine physical examinations and blood tests were performed to assess and maintain their health status.
NB-UVR LED devicesIn this study, LED devices (Nichia Corporation, Tokushima, Japan) with peak wavelengths of 310, 320, and 330 nm were used for irradiation. All the LED devices had a half-bandwidth of 10 nm (Supplementary Fig. 1). The intensity of the irradiated surface was standardized to 6.5 mW/cm2.
Clinical trial of NB-UVR therapy in dogs with spontaneous ADThe skin lesions of the six client-owned dogs with spontaneous AD were treated with 310 nm or 320 nm NB-UVR irradiation (Supplementary Fig. 2). A lesion treated with a topical betamethasone valerate ointment (RINDERON-VG Ointment; Shionogi, Osaka, Japan) served as the positive control, while an untreated lesion was used as the negative control (Supplementary Table 3). Each dog had an equal number of lesions that were selected for NB-UVR treatment. Among these, one lesion per dog was designated as the positive control treatment and the other was designated as the negative control (Supplementary Table 3).
The exclusion criteria were active systemic treatments (e.g., cyclosporine, antibiotics, and antifungal drugs), a change in shampoo or diet within ≤2 weeks before the trial, receiving vaccinations within ≤2 weeks before the trial, and a history of malignant diseases. Treatments were administered weekly or biweekly, based on the owners’ schedules. We strictly selected cases that met the canine AD diagnosis and exclusion criteria, and owner consent was obtained for enrollment in the study. A preliminary statistical analysis was performed after five animals were enrolled. When a trend was observed, an additional animal was included, resulting in six animals for evaluation. There were no dropouts during the study period.
Preliminary experiments with healthy beagles established that the MED range for NB-UVR irradiation was 400–1,500 mJ/cm2 at 310 nm (n=6) and >1,500 mJ/cm2 at 320 nm (n=3) (Supplementary Table 2). Based on the MED at 310 nm NB-UVR, an initial dose of 300 mJ/cm2 was selected for the clinical trials. This dose was then used as the sub-MED in subsequent in vivo and in vitro experiments. To standardize the initial irradiation dose, the MED was not assessed in individual canine AD cases. The dose was increased by 100 mJ/cm2 at each follow-up visit. The dosage of betamethasone ointment was 0.05–0.1 g per lesion (approximately one fingertip unit).
Lesion severity was evaluated at each visit using the Canine Atopic Dermatitis Extent and Severity Index (CADESI-4) lesion grading atlas [20]. Four clinical signs (erythema, alopecia, excoriation, and lichenification) were scored on a 6-point scale (0=none, 1–2=mild, 3=moderate, and 4–5=severe). The total lesion score was calculated by summing the scores for each sign (range: 0–20 points). The average score was used for dogs with more than two NB-UVR-treated lesions. The percentage change in the total lesion score relative to the baseline (week 0) was calculated. The scores for NB-UVR treatments (310 and 320 nm) were evaluated in a double-blind manner.
IDTTo assess the effect of NB-UVR on anaphylactoid reactions, three beagles underwent IDT under sedation with medetomidine (0.02 mg/kg, IM; Dorbene Vet, Kyoritsu Seiyaku, Tokyo, Japan) and midazolam (0.3 mg/kg, IM; Dormicum injection 10 mg, Astellas Pharma Inc., Tokyo, Japan). Prior to IDT, NB-UVR (310, 320, and 330 nm) was applied to the same hair-clipped skin sites at a dose of 300 mJ/cm2 once daily for four consecutive days. On the day after the final irradiation, IDT was performed in triplicate at 12 sites: three NB-UVR-irradiated sites and one unirradiated site on the right lateral, back, and left lateral regions.
Polyoxyethylene-hydrogenated castor oil 60 (HCO-60; Kao Chemicals, Tokyo, Japan) was used as an indicator of MC degranulation. Concentrations of 1 × 10−6 v/v (0.001%) and 1 × 10−5 v/v (0.01%) were selected based on a previously established threshold for MC degranulation in healthy dogs [21]. Histamine (1 × 10−5 w/v; Sigma Aldrich, St. Louis, MO, USA) was used as the positive control, while saline served as the negative control. The reagents (50 µl) were serially diluted in saline to a 1:10 ratio and injected intracutaneously.
Photographs of the pre- and post-IDT sites were taken to assess changes in the redness values (RV) and maximum diameter (D) of the wheal-and-flare formations, which were measured 15 min after injection. RV was analyzed using computer software (GNU Image Manipulation Program, https://www.gimp.org/) by calculating the mean red (R) and green (G) intensities (range: 0–255) using a color picker in the RGB mode. RV was determined using the following formula:
RV=(R−G)/(R+G)Changes in RV were expressed as the ratio of the post-IDT RV to the pre-IDT RV. The IDT score was calculated as follows:
IDT score=D×RV changeBoth the D and RV were measured in a blinded manner. Following this procedure, atipamezole (0.08 mg/kg; Atipame injection, Kyoritsu Seiyaku) was administered intramuscularly for reverse sedation.
Skin biopsyNB-UVR (310, 320, and 330 nm) was applied to the back region of three beagles at a sub-MED of 300 mJ/cm2 once daily for four consecutive days, following the same protocol described for IDT. A higher NB-UVR dose (310 nm, 1,500 mJ/cm2) was applied as an inflammatory control one day before skin biopsy.
Skin biopsies were performed with the dogs under sedation with medetomidine and midazolam. Tissue samples were collected from the irradiated sites using an 8-mm biopsy punch. Each sample was divided into two halves: one was fixed in 9 ml formalin, methanol, and picric acid-based fixative (Gfix; Genostaff, Tokyo, Japan) for histopathological analysis, and the other was preserved in RNAlater solution (Thermo Fisher Scientific, Tokyo, Japan) for the RNA-seq.
Histopathological analysisThe fixed tissues were embedded in paraffin, sectioned at 5-µm thickness, and stained with hematoxylin and eosin (HE) or toluidine blue (TB, pH 4.1; Muto Pure Chemicals, Tokyo, Japan) for analysis of sunburn cells and MCs. Epidermal sunburn cells were identified in HE-stained sections, whereas MCs were counted in TB-stained sections. The focus depth was set at 0.2 mm, covering the superficial dermis to the upper reticular dermis. The average number of MCs per section was calculated from the counts obtained from four sections per sample. All the histopathological evaluations were performed in a blinded manner.
The RNA-seq analysisThe RNA-seq analysis was performed using Takara Bio (Shiga, Japan). Total RNA was extracted using the RNAiso Plus kit (Takara Bio) and purified using the NucleoSpin RNA Clean-up XS kit (Takara Bio). The RNA concentration and quality were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific) and an Agilent 2200 TapeStation System (Agilent Technologies, Santa Clara, CA, USA), respectively. Sequence libraries were prepared using a TruSeq Stranded mRNA Sample Prep Kit (Illumina, San Diego, CA, USA) and indexed with IDT for Illumina TruSeq UD Indexes (Illumina) on an Agilent XT-Auto System (Agilent Technologies). For each sample, 2 µg RNA was used for PolyA+ RNA isolation. Adaptor-ligated double-stranded DNA was amplified using 15 PCR cycles. Sequencing was performed on the NovaSeq 6000 system (Illumina) using the NovaSeq 6000 S4 Reagent Kit and NovaSeq Xp 4-Lane Kit for 150-bp paired-end reads. The RNA-seq reads were mapped to the Canis lupus familiaris reference genome (NCBI Taxonomy ID: 9615, CanFam3.1, assembly, April 2018) using STAR v2.5.2b and Genedata Profiler Genome v10.1.15a (Genedata, Basel, Switzerland).
In vitro cell viability and apoptosis assayThe canine cutaneous MC tumor cell line HRMC [22] was used for in vitro experiments. The cells were cultured in RPMI 1640 medium (Wako Pure Chemical Industries, Osaka, Japan) supplemented with 10% fetal bovine serum (Biowest, France) and 1% penicillin/streptomycin (Sigma-Aldrich). The cultures were maintained at 37°C in a humidified atmosphere containing 5% CO2.
The effect of NB-UVR irradiation on HRMC viability was assessed using the WST assay with Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan). The cells were seeded at a density of 1 × 106 cells/ml in 96-well plates and irradiated with NB-UVR at various doses (100–600 mJ/cm2) and wavelengths (310, 320, and 330 nm). After 24 hr of incubation, the CCK-8 reagent was added to each well, followed by a 4-hr incubation. The absorbance was measured at 450 nm using a microplate reader (Power Scan HT, DS Pharma Biomedical, Tokyo, Japan).
Apoptosis was evaluated using DAPI staining, following a previously described protocol [23]. HRMCs (1 × 106 cells/ml) were irradiated with NB-UVR (310, 320, or 330 nm) at 300 mJ/cm2. After 24 hr, cells were adhered to glass slides using LBC Prep 2 (Muto Pure Chemicals), fixed with 4% paraformaldehyde, and permeabilized with 0.2% Triton X-100 (Kanto Chemical, Tokyo, Japan). DAPI solution (VECTASHIELD Mounting Medium with DAPI; Vector Laboratories, Burlingame, CA, USA) was then applied to slides. Apoptotic cells were identified based on nuclear fragmentation and chromatin condensation, using a fluorescence microscope (AX80N-05; Olympus, Tokyo, Japan). Five fields were analyzed for each treatment condition and the experiments were performed in triplicate.
Statistical analysisOne-way repeated-measures analysis of variance was used to evaluate the total lesion scores, IDT scores, MC numbers, in vitro cell viability, and apoptosis rates for each treatment. Significant differences were further analyzed using Dunnett’s test to compare the NB-UVR-irradiated groups with the non-irradiated controls. Comparisons among the NB-UVR wavelengths (310, 320, and 330 nm) were conducted using Tukey’s multiple comparison test. The Wilcoxon rank-sum test was utilized for pairwise comparisons of each lesion sub-score (erythema, alopecia, excoriation, and lichenification) between pre-irradiation and after the 4th irradiation. A P-value of <0.05 was considered statistically significant. Statistical analyses were performed using the StatMate V software (ATMS, Tokyo, Japan).
Low read counts (<100 total counts per gene) were excluded from the RNA-seq data. Normalization and differential expression analysis were performed using the Tag Count Comparison (TCC) baySeq R package [24, 25]. Using the normalization factors obtained from the TCC, baySeq was employed to calculate the probability of each expression pattern and identify the most probable one. Differentially expressed genes (DEGs) were identified using a false discovery rate (FDR) threshold of q<0.1. The dataset was categorized based on the UVR treatment (non-irradiated control and UVR irradiation) and experimental conditions (wavelength and dose) to identify DEGs associated with UVR irradiation. Groups 1–5 corresponded to the non-irradiated control (G1) and NB-UVR irradiation at different wavelengths and doses: G2 (300 mJ/cm2 and 310 nm), G3 (300 mJ/cm2 and 320 nm), G4 (300 mJ/cm2 and 330 nm), and G5 (1,500 mJ/cm2 and 310 nm). The probabilities of different expression patterns were classified using baySeq as follows: non-DEG (no differentially expressed genes), DEG_G1 to DEG_G5 (differentially expressed genes in groups 1–5), and DEG_all (differential expression across all groups). If a pattern matched, the order of groups, indicating whether expression levels were higher or lower than those in other groups, was determined based on the normalized expression counts (e.g., G1 >others, G2 >others, G5 >G3 >G4 >G1 >G2, and G5< G3< G4< G2< G1).
The identified DEGs were visualized using Heatmapper [26] to illustrate the relationship between gene expression levels and treatment groups. For the hierarchical clustering analysis, the average linkage method and Kendall’s tau distance metric were applied to both rows (DEGs) and columns (dogs and treatments), and the dendrogram was displayed along with a heatmap. A heatmap was generated by converting the TCC normalized gene expression counts to Z-scores using the scaling option with the default settings.
Gene ontology (GO) enrichment analysis (Biological Process Level 3) and KEGG pathway analyses were performed using DAVID Bioinformatics Resources 2021 (https://david.ncifcrf.gov/), with an FDR threshold of q<0.1. The RNA-seq data were deposited in the DDBJ BioProject database under BioProject accession number PRJDB20532.
By the fourth treatment session, significant differences in the total lesion scores were observed among the groups. The 320 nm irradiation group showed significant improvement compared to the untreated control group (n=6, P<0.05; Fig. 1A, 1B). A comparison of skin lesion sub-scores (erythema, alopecia, excoriation, and lichenification) between baseline (before the trial) and the 4th treatment revealed significant reductions (Supplementary Table 4). Specifically, the alopecia sub-score significantly decreased in the 310 nm irradiation, 320 nm irradiation, and betamethasone treatment groups (P<0.05). A significant reduction in the excoriation sub-score was observed in the 310 nm irradiation and betamethasone treatment groups (P<0.01), whereas the lichenification sub-score showed a significant reduction in the 310 nm irradiation group (P<0.01) and 320 nm irradiation group (P<0.05). No adverse clinical reactions were observed in any dogs during the study period.
Effect of Narrowband ultraviolet radiation (NB-UVR) on skin lesions in canine atopic dermatitis. (A) Changes in skin lesion scores after the fourth treatment with NB-UVR, topical ointment, or no treatment. The NB-UVRs doses were adjusted to 600 mJ/cm2. * indicates a significant difference compared to the untreated control (n=6, P<0.05). Data are presented as mean ± standard deviation, (B) Representative clinical findings from a dog treated with 320 nm NB-UVR. Before treatment, alopecia, lichenification, and erythema were observed on the cervical skin. Four weeks after treatment, hair regrowth and reduced erythema were evident.
The IDT scores for the reactions induced by HCO-60 (1 × 106 v/v) and histamine differed significantly from those of the non-irradiated controls at all NB-UVR wavelengths (n=3, P<0.01). Furthermore, the IDT score decreased significantly (P<0.05) at 320 nm in response to 1 × 105 v/v of HCO-60 (Fig. 2).
Effect of NB-UVR on polyoxyethylene hydrogenated castor oil 60 (HCO-60)- and histamine-induced wheal-and-flare reactions in beagles. NB-UVR was applied at a dose of 300 mJ/cm2. ** indicates a significant difference compared to the non-irradiated control *(P<0.05) and **(P<0.01), n=3. Data are presented as mean ± standard deviation.
The sunburn cells are keratinocytes that undergo apoptosis owing to physiological UVR damage [27]. No sunburned cells were detected in any of the groups that received 300 mJ/cm2 NB-UVR irradiation. However, irradiation at 310 nm at a higher dose (1,500 mJ/cm2) induced visible sunburn in the cells (Fig. 3A). MC counts in skin sections were significantly reduced in the 310 nm NB-UVR groups compared to the control group (n=3, P<0.05; Fig. 3B and 3C).
Effect of NB-UVR on histopathology and skin mast cell number in beagles. (A) Histopathological findings of NB-UVR irradiated canine skin. Inflammatory cell infiltration of the superficial dermis and apoptotic keratinocytes (arrow) observed in 1,500 mJ/cm2 310-nm NB-UVR irradiated canine skin. (HE staining, ×100), (B) Skin mast cell number in skin irradiated with NB-UVR. An asterisk indicates a significant difference compared to non-irradiated controls (n=3, P<0.05). Data are presented as mean ± standard deviation. (C) The toluidine blue-staining skin sections. Red arrows indicate dermal mast cells.
Irradiation with 310 and 320 nm NB-UVR reduced HRMC viability in a dose-dependent manner, whereas no significant effects were observed with 330 nm NB-UVR (Fig. 4A). At doses ranging from 200 to 600 mJ/cm2, 310 nm NB-UVR significantly reduced cell viability compared to 320 and 330 nm NB-UVR (P<0.01 and P<0.001, respectively). Additionally, cells treated with 310 nm NB-UVR exhibited a significantly increased rate of apoptosis (P<0.01; Fig. 4B).
Effects of NB-UVRs on cell viability and apoptosis in HRMC cells. (A) Cell viability assessed using the WST assay, (B) Rate of apoptotic cells determined using DAPI staining. The NB-UVR dose was set at 300 mJ/cm2 for apoptosis assessment. * (P<0.05) and ** (P<0.01) indicate a significant difference compared to non-irradiated cells. §(P<0.01) and §§(P<0.001) indicates a significant difference compared to cells treated with 310 nm NB-UVR. Experiments were performed in triplicate. Data are presented as mean ± standard deviation.
In total, 15,399 expressed genes were subjected to filtering. Ultimately, 39 non-DEGs and 88 statistically significant DEGs (n=3, q<0.1) were identified using TCC-baySeq analysis (Supplementary Table 5). Among the 88 DEGs, the number of genes for each NB-UVR-irradiated group comparison (G3 >others, G5 >others, others >G5, others >G2, and others >G1) were 71, 13, 2, 1, and 1, respectively.
A summary of the functional enrichment analysis of the DEGs (q<0.1) is presented in Table 1. Irradiation with 320 nm NB-UVR at 300 mJ/cm2 was associated with enrichment in muscle components (M band, sarcoplasmic reticulum, and Z-disc) and functions (contraction and cytoskeletal regulation), carbohydrate metabolism (glycolysis/gluconeogenesis, fructose 1,6-bisphosphate metabolic processes, pentose phosphate pathway, and fructose and mannose metabolism), biosynthesis of amino acids, signaling pathways (glucagon signaling and adrenergic signaling in cardiomyocytes), and neural components (axons). In contrast, irradiation with 310 nm NB-UVR at 1,500 mJ/cm2 enriched only the innate immune response.
| Orderings of Narrowband-UVR Irradiationa) | Category | Term (ID) | Count/List total | % | P-value | Gene symbol | Pop hits/Pop total | FDR |
|---|---|---|---|---|---|---|---|---|
| 300 mJ/cm2,320 nm>others | GOTERM_BP_DIRECT | Muscle contraction (GO:0006936) | 4/58 | 6.0 | 2.36E-04 | tmod4, lmod3, clcn1, tnni2 | 40/18940 | 6.04E-02 |
| GOTERM_BP_DIRECT | Canonical glycolysis (GO:0061621) | 3/58 | 4.5 | 5.76E-04 | pgam2, eno3, pfkm | 12/18940 | 6.75E-02 | |
| GOTERM_BP_DIRECT | Fructose 1,6-bisphosphate metabolic process (GO:0030388) | 3/58 | 4.5 | 7.91E-04 | fbp2, aldoa, -pfkm | 14/18940 | 6.75E-02 | |
| GOTERM_CC_DIRECT | M band (GO:0031430) | 4/60 | 6.0 | 1.25E-05 | myom3, lmod3, myom1, aldoa | 16/20408 | 1.09E-03 | |
| GOTERM_CC_DIRECT | Sarcoplasmic reticulum (GO:0016529) | 4/60 | 6.0 | 8.81E-05 | jph1, s100a1, trdn, tmem38a | 30/20408 | 3.83E-03 | |
| GOTERM_CC_DIRECT | Z disc (GO:0030018) | 5/60 | 7.5 | 1.65E-04 | ldb3, jph1, hrc, fbp2, trim54 | 95/20408 | 4.78E-03 | |
| GOTERM_CC_DIRECT | Axon (GO:0030424) | 5/60 | 7.5 | 4.27E-03 | myot, scn1b, slc8a3, kcnc1, tnni2 | 22/204088 | 9.29E-02 | |
| KEGG_PATHWAY | Cytoskeleton in muscle cells (cfa04820) | 14/37 | 20.9 | 2.46E-12 | ldb3, obscn, myot, trim55, ankrd23, myom3, tmod4, lmod3, mybpc2, myom1, eno3, trim54, capn3, tnni2 | 234/8989 | 2.48E-10 | |
| KEGG_PATHWAY | Glucagon signaling pathway (cfa04922) | 6/37 | 9.0 | 6.11E-05 | pgam2, phkg1, fbp2, adcy2, prkag3, pfkm | 107/8989 | 3.09E-03 | |
| KEGG_PATHWAY | Glycolysis/Gluconeogenesis (cfa00010) | 14/37 | 7.5 | 3.81E-04 | pgam2, eno3, fbp2, aldoa, pfkm | 87/8989 | 1.28E-02 | |
| KEGG_PATHWAY | Carbon metabolism (cfa01200) | 14/37 | 7.5 | 1.77E-03 | pgam2, eno3, fbp2, aldoa, pfkm | 131/8989 | 4.46E-02 | |
| KEGG_PATHWAY | Adrenergic signaling in cardiomyocytes (cfa04261) | 14/37 | 7.5 | 2.97E-03 | popdc3, scn1b, slc8a3, ppp1r1a, adcy2 | 151/8989 | 5.99E-02 | |
| KEGG_PATHWAY | Biosynthesis of amino acids (cfa01230) | 4/37 | 6.0 | 5.80E-03 | pgam2, eno3, aldoa, pfkm | 92/8989 | 9.62E-02 | |
| KEGG_PATHWAY | Pentose phosphate pathway (cfa00030) | 3/37 | 4.5 | 7.17E-03 | fbp2, aldoa, pfkm | 32/8989 | 9.62E-02 | |
| KEGG_PATHWAY | Fructose and mannose metabolism (cfa00051) | 3/37 | 4.5 | 7.62E-03 | fbp2, aldoa, pfkm | 33/8989 | 9.62E-02 | |
| 1500 mJ/cm2, 310 nm >others, and others > 1500 mJ/cm2 310 nm | GOTERM_BP_DIRECT | Innate immune response (GO:0045087) | 5/15 | 33.3 | 2.39E-05 | ptx3, bst2, mmp3, oas1, oasl | 243/18940 | 1.96E-03 |
a) The orderings were determined by RNA-seq analysis using TCC baySeq. TCC: tag count comparison; UVR: ultraviolet radiation; FDR: false discovery rate.
Hierarchical clustering analysis, as shown by the heatmap and dendrogram (rows) of DEG expression levels, revealed that dogs subjected to irradiation with 320 nm NB-UVR at 300 mJ/cm2 exhibited similar heatmap patterns and were included within the same cluster, whereas those irradiated with 310 nm NB-UVR at 1,500 mJ/cm2 also showed similar heatmap patterns but were represented in distinct clusters, appearing adjacent to the dendrogram (Fig. 5). Among the 71 DEGs identified in the G3 >others group, 68 DEGs exhibited more than a 2-fold change (median: 8.3-fold; range: 1.7 to 1,270.0-fold), while for the 13 DEGs in the G5 >others group, 12 DEGs showed changes (median: 3.2-fold; range: 1.5 to 41.7-fold) (Supplementary Table 5). Regarding the dendrogram (columns) for dogs and treatments, all dogs irradiated with 320 nm NB-UVR at 300 mJ/cm2 clustered together and separated from the cluster containing the control groups. In contrast, dogs treated with 310 nm NB-UVR at 1,500 mJ/cm2 appeared in an adjacent column yet belonged to a different cluster (Fig. 5).
Heatmap of differentially expressed genes (DEGs) identified after NB-UVR irradiation. A heatmap based on Z-scores was generated to visualize the results of the hierarchical clustering analysis performed on dogs subjected to NB-UVR treatment (columns) and DEGs (rows). The DEGs were identified using TCC-baySeq analysis (n=3, q<0.1). DEGs detected by functional enrichment analysis (DAVID) are highlighted in bold.
This study investigated the effects of the sub-MED of NB-UVR on canine skin to clarify the biological impact of low-dose NB-UVR phototherapy, a therapeutic approach with many unresolved aspects. First, we examined the therapeutic potential of NB-UVR for spontaneous canine AD. Unexpectedly, NB-UVR irradiation at 320 nm, even when administered below the MED, resulted in a significant reduction in total lesion scores. This effect was superior to that of 310 nm NB-UVR, which was administered at doses (600 mJ/cm2) approaching previously reported effective levels (750 mJ/cm2 at 308 nm NB-UVR) [18]. Next, we investigated the mechanisms underlying the sub-MED of NB-UVR at different wavelengths (310, 320, and 330 nm) using experimental beagles. The results demonstrated that sub-MED NB-UVR suppressed skin anaphylactoid reactions (at all wavelengths), reduced skin MC counts (at 310 nm), and altered the expression of non-inflammatory genes (DEGs) as indicated using the RNA-seq analysis (at 320 nm). Furthermore, in vitro studies revealed that sub-MED NB-UVR reduced MC viability (310 and 320 nm) and induced apoptosis (310 nm). These findings suggest that sub-MED NB-UVR exhibits distinct wavelength-specific therapeutic mechanisms; irradiation at 310 nm induces MC apoptosis, whereas irradiation at 320 nm activates a non-inflammatory pathway. The novel non-inflammatory effect observed with 320 nm irradiation holds promise for the development of innovative therapeutic strategies in human medicine.
Among the various guidelines for treating spontaneous AD in dogs [13], those for NB-UVR therapy are limited. Previous studies have reported MED values, with 311 nm NB-UVR requiring ≥432 mJ/cm2 [16] and 308 nm NB-UVR requiring ≥560 mJ/cm2 [17], which aligns with our preliminary MED determination experiments at 310 nm NB-UVR (≥400 mJ/cm2; Supplementary Table 2). Additionally, 308 nm NB-UVR has been shown to improve skin barrier function, clinical symptoms, and skin microbiome dysbiosis of canine AD [18], and induce apoptosis and reduce INF-γ and TNF-α levels in a canine dermatitis model stimulated by dinitrochlorobenzene [17], suggesting that UVR effectively modulates skin surface and penetrates into the canine dermis to exert its effects. Despite no significant difference in the total lesion score (Fig. 1), 310 nm irradiation significantly improved the alopecia, excoriation, and lichenification sub-scores in canine AD (Supplementary Table 4). Even at sub-MED doses, we demonstrated suppressed anaphylactoid reactions, reduced cutaneous MC numbers, and induced apoptosis in cultured MCs. These findings support previously reported efficacy of NB-UVR therapy in dogs.
In our clinical trial, a significant reduction in the total lesion score, particularly in the alopecia and lichenification sub-scores, was observed in the low-dose 320 nm NB-UVR (Fig. 1 and Supplementary Table 4). The RNA-seq analysis revealed that sub-MED 320 nm NB-UVR irradiation upregulated the expression of genes associated with muscle function (GO:0006936, GO:0031430, GO:0016529, GO:0030018, cfa04820, and cfa04261) (Table 1). The muscular components of the skin include the arrector pili muscles (APMs) and the cutaneous muscles. The APM is a smooth muscle located in the upper and middle dermis that connects the epidermal basement membrane to the bulge region of hair follicles (Fig. 3A), whereas the cutaneous muscles are located beneath the subcutaneous adipose tissue. Although further verification is necessary, the fact that NB-UVR does not penetrate beyond the subcutaneous fat layer suggests that its effects may have been exerted on the APMs. Schwartz et al. reported that sympathetic nerves innervate both the APMs and hair follicle stem cells (HFSCs) [28]. Piloerection is triggered by APM contraction and plays a role in maintaining sympathetic innervation of HFSCs. The β2-adrenergic receptor (ADRB2) in HFSCs directly responds to norepinephrine and cold signals from the nervous system, activating muscle contraction and initiating a new hair cycle via the GPCR/Gαs/AC/cAMP/CREB pathway [29]. Our results identified genes related to adrenergic signaling in the cardiomyocyte pathway (cfa04261) and enzymes involved in glycolysis and gluconeogenesis (cfa00010). These pathways/enzymes have been reported to be associated with HFSC activation [29], suggesting that 320 nm NB-UVR irradiation may induce changes in HFSCs similar to those triggered by cold stimulation.
The transcription factor-encoding gene Mef2c and the calcium signaling-related genes Hrc and Trdn were among the DEGs upregulated by 320 nm irradiation (Supplementary Table 5). Myocyte Enhancer Factor 2C (MEF2C) is a critical regulator of muscle and nervous system development and function [30, 31]. MEF2C directly upregulates Hrc expression, which encodes a histidine-rich calcium-binding protein (HRCBP). HRCBP localizes to calciosomes within arterial smooth muscle cells [32] and interacts with triadin in the sarcoplasmic reticulum [33]. Both HRCBP and triadin play essential roles in calcium signaling pathways (cfa04020), influencing muscle contraction by regulating intracellular calcium concentrations. The therapeutic effect of the 320 nm wavelength on canine AD observed in the clinical trial may be associated with the upregulation of these genes, which are involved in hair follicle development and muscle function.
Several studies have investigated the effects of UV-induced immunomodulation in MCs [4,5,6,7]. Contrary to our findings, previous studies have reported an increase in cutaneous MC numbers and histamine release following UVR exposure [5,6,7, 10]. These studies primarily involved irradiation at doses exceeding the MED, and the observed effects were attributed to inflammation induced via the keratinocyte damage pathway. Similarly, in our study, irradiation at a dose above the MED (1,500 mJ/cm2, 310 nm) resulted in changes in the innate immune response (GO:0045087) (Table 1). In contrast, we were unable to identify DEGs using TCC-bayseq following 310 nm sub-MED NB-UVR irradiation. We propose that this may be attributed to the experimental setup involving healthy skin, which is less likely to harbor cells actively undergoing apoptosis upon UV absorption than damaged or diseased tissues [17]. It is plausible that basal gene expression levels in these quiescent cells were below the threshold required to identify irradiation-induced changes.
In this study, basic experiments using experimental beagles were conducted with a sub-MED of the NB-UVR. Most previous studies have utilized doses above the MED, and research on sub-MED exposure remains limited. Danno et al. [8] demonstrated that UVR exposure at sub-MED doses suppressed IDT reactions and MC degranulation in mice. Similarly, Byne et al. [9] reported dose-dependent suppression of CHS. Consistently, our study showed that sub-MED (300 mJ/cm2) of NB-UVR (310, 320, and 330 nm) exhibited anti-anaphylactic effects in dogs (Fig. 2). Furthermore, Danno et al. [8] suggested that UVR with a peak emission at 305 nm (wavelength range: 280–370 nm) exerts a dose-dependent dual effect on MCs, where doses >200 mJ/cm2 induce inflammation, whereas MED alters the MC/vasoactive amine system and suppresses ear swelling in response to degranulators. Byne et al. [9] also reported different dose-response effects for UVR (320–400 nm) compared to UVR (290–320 nm). Similarly, our results indicated that different wavelengths of sub-MED NB-UVR elicited distinct MC responses to UVR exposure. The device used in this study had a narrow half-bandwidth of 10 nm, enabling more precise analysis of specific wavelength ranges compared to previous studies on sub-MED NB-UVR.
Canine AD is defined as “a hereditary, typically pruritic and predominantly T-cell driven inflammatory skin disease involving interplay between skin barrier abnormalities, allergen sensitization and microbial dysbiosis” [34]. Given its established clinical and immunological similarities to human AD, its spontaneous nature, and the shared living environments between dogs and their owners, canine AD offers a unique opportunity for preclinical research [11,12,13,14, 19, 34]. The availability of advanced veterinary medicine, including novel JAK inhibitors and anti-IL-31 antibodies, further positions canine AD as a suitable animal model for human AD. Dogs are entirely covered in fur and have thinner skin than humans. This makes them valuable models for studying pathologies in hair-bearing regions such as the head, which represent relatively small body surface areas in humans. Our results confirmed the validity of our experimental approach, indicating that the NB-UVR LED device used in our experiment functions similarly to conventional UVR irradiators. By focusing on the challenges in medical research stemming from the physiological differences between dogs and humans, we successfully clarified the in vivo biological responses to sub-MED 320 nm NB-UVR in dogs. This study generated crucial preclinical data relevant to humans.
This study had several limitations. This was a pilot study with a small sample size, indicating the need for a large-scale comparative controlled study. Additionally, because we focused on setting the sub-MED at 310 nm, the effect at 320 nm may not have been sufficiently evaluated. Additionally, the RNA-seq analysis was performed on whole-thickness skin biopsies containing mixed cell populations, which may have obscured cell-specific gene expression. Although the observed transcriptomic changes suggested biological effects, they were not validated at the protein level. Further studies are required to clarify the mechanisms of sub-MED NB-UVR by incorporating cell-type-specific analyses and protein-level validation.
Overall, this study demonstrated the efficacy and underlying mechanisms of sub-MED NB-UVR exposure, highlighting dose- and wavelength-specific effects. Low-dose NB-UVR at 320 nm improved the clinical signs without inducing erythema, suggesting a potentially low-risk therapeutic approach. Considering the similarities between canine and human AD, these findings may have translational relevance. Irradiation at 320 nm is associated with non-inflammatory gene expression involving GPCR signaling and muscle-related pathways, possibly linked to hair follicle stem cells. These responses may indicate a role in skin regeneration and support the future development of novel non-erythemogenic phototherapies. Although the current NB-UVR therapies often require doses exceeding the MED, our results suggest that sub-MED NB-UVR may offer clinical benefits with fewer adverse effects. Overall, these findings provide preliminary support for further investigation of low-dose NB-UVR as a therapeutic option for human dermatology.
YM and SK1 prepared the original draft. AT and SK2 reviewed and edited the manuscript. YM, SK1, and HO performed the in vivo experiments, whereas MK and MN performed the in vitro experiments. SM and TT analyzed the RNA-seq data. YO and YF developed light-emitting diode narrowband ultraviolet radiation devices. KN edited the manuscript for improved clarity and grammar. SA reviewed and interpreted the histopathological findings. SA supervised the medical statements. AT and SK2 secured funding. SK2 contributed to the conceptualization and methodology of this study.
This study was supported by JSPS KAKENHI grant number JP17K18189 (S.K.). This research was funded by the Nichia Corporation (A.T. and S.K.). The LED device for NB-UVR exposure has already been submitted to Japanese and international patent offices (Number 2022-169231 and WO2020/090919, respectively).
The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.
This research was initially made available as a preprint of bioRxiv prior to submission (doi: https://doi.org/10.1101/2025.01.01.631029). The preprint has been revised before submission for publication.
The authors sincerely thank Dr. Keitaro Ohmori from the Cooperative Division of Veterinary Sciences, Graduate School of Agriculture, Tokyo University of Agriculture and Technology for generously providing the HRMC cell line. We thank Dr. Azusa Ogita and Dr. Keigo Ito of the Division of Dermatology and Dermatopathology, Nippon Medical School, Musashi Kosugi Hospital, and Dr. Hidenori Matsuda of the Shibuya Aesthetic Surgery Clinic for sharing their knowledge and skills in dermatology practice. We thank Hitomi Endo for her assistance with the animal experiments.
