Human mesenchymal stem cells are resistant to UV-B irradiation

Albeit being an effective therapy for various cutaneous conditions, UV-B irradiation can cause severe skin damage. While multipotent mesenchymal stem cells (MSCs) may aid the regeneration of UV-B-induced skin injuries, the influence of UV-B irradiation on MSCs remains widely unknown. Here, we show that human MSCs are relatively resistant to UV-B irradiation compared to dermal fibroblasts. MSCs exhibited higher clonogenic survival, proliferative activity and viability than dermal fibroblasts after exposure to UV-B irradiation. Cellular adhesion, morphology and expression of characteristic surface marker patterns remained largely unaffected in UV-irradiated MSCs. The differentiation ability along the adipogenic, osteogenic and chondrogenic lineages was preserved after UV-B treatment. However, UV-B radiation resulted in a reduced ability of MSCs and dermal fibroblasts to migrate. MSCs exhibited low apoptosis rates after UV-B irradiation and repaired UV-B-induced cyclobutane pyrimidine dimers more efficiently than dermal fibroblasts. UV-B irradiation led to prolonged p53 protein stability and increased p21 protein expression resulting in a prolonged G2 arrest and senescence induction in MSCs. The observed resistance may contribute to the ability of these multipotent cells to aid the regeneration of UV-B-induced skin injuries.

UV-B treatment leads to heterogeneous results regarding MSCs adhesion ability. MSC adherence was examined over a period of 24 hours after UV-B exposure. While there was no delay in cellular attachment of MSCs, number of attached cells differed between unirradiated and UV-B-irradiated cells in MSC1 and MSC3 24 hours after UV-B irradiation. MSC1 cells were shown to exhibit lower adhesion rates 24 hours after low-dose (25 mJ/cm 2 ) but not after high-dose UV irradiation (100 mJ/cm 2 ) (P < 0.05). In contrast, irradiation with 100 mJ/cm 2 resulted in a significant reduction of adhesion in MSC3 (P < 0.01), whereas low-dose irradiation with 25 mJ/cm 2 led to comparable adhesion rates between irradiated and untreated cells. MSC2 showed no changes in their cellular attachment rates 24 hours after UV irradiation (P = 0.29 for 25 mJ/cm 2 , P = 0.95 for 100 mJ/cm 2 ) (Fig. 2a). Both low-dose (25 mJ/cm 2 ) and high-dose (100 mJ/cm 2 ) UV-B irradiation did not reduce the adhesion ability of HS68 fibroblasts (P = 0.25 for 25 mJ/cm 2 , P = 0.52 for 100 mJ/cm 2 ). using doses up to 1500 mJ/cm 2 . *P < 0.05, **P < 0.01, ***P < 0.001. Two-sided Student's t-tests at 200 mJ/ cm 2 (clonogenicity and proliferation assays) and 1500 mJ/cm 2 (viability assays) were used. Mean ± standard deviation is shown, n = 3.

MSC surface marker expression and morphology are unaffected by UV-B irradiation.
Surface marker expression of MSCs 96 hours after UV-B irradiation was examined by flow cytometry. Both 25 mJ/cm 2 and 100 mJ/cm 2 did not affect the expression of positive stem cell surface markers CD73, CD90 and CD105 in all tested MSC samples (Fig. 3a). Similarly, the lack of expression of the hematopoietic markers CD14, CD20, CD34 and CD45 remained unchanged after UV-B exposure.

Low-dose UV-B irradiation leads to G2/M arrest in MSCs.
Flow cytometry analyses were carried out to analyze cell cycle distribution after UV-B exposure. Low-dose irradiation caused an accumulation of cells in the G2/M phase 24 hours after treatment, which persisted at later time points (Supplementary Fig. 2). The cell cycle distribution of MSCs after high-dose irradiation appeared more heterogeneous and after 96 hours, only MSC1 and MSC3 exhibited an increase in their G2 phase population (P < 0.001), whereas MSC2 revealed a small decrease (P < 0.05). HS68 fibroblasts clearly showed a dose-dependent G2/M phase accumulation upon UV irradiation with 100 mJ/cm 2 . This pronounced and lasting G2/M arrest involves more than 50% cells of the population compared to 23.3% in untreated controls (P < 0.001), while low-dose UV-B only causes a minor increase.

UV-B irradiation induces low apoptosis rates in MSCs.
Sub-G1 population and caspase-3 activation were measured to quantify apoptosis induction after UV-B treatment. Overall apoptosis rates remained low in all analyzed MSC samples with levels below 10% (Fig. 5a). In contrast, dermal fibroblasts exhibited increased apoptosis rates after both low-dose (25 mJ/cm 2 ) and high-dose (100 mJ/cm 2 ) UV-B irradiation, and more than 50% of HS68 cells were apoptotic 96 hours after 100 mJ/cm 2 as determined by caspase-3 activation (P < 0.05).
MSCs exhibit a more efficient CPD repair than dermal fibroblasts. Repair of UV-B-induced CPDs was investigated using ELISA analyses. CPD levels were found elevated already 30 minutes both after 25 mJ/ cm 2 and 100 mJ/cm 2 UV irradiation in MSC1 and HS68 and remained stably elevated until the 6-hour timepoint (P < 0.001) (Fig. 6a). 24 hours after UV-B irradiation, CPD levels were significantly increased compared to untreated controls both for MSC1 and HS68 cells (P < 0.001). However, MSCs exhibited about 25% lower CPD levels than HS68 fibroblasts 24 hours after 100 mJ/cm 2 .

UV-B irradiation increases expression of p53 and p21 in MSCs.
Western blot analyses revealed p53 stabilization and increased expression of p21 in MSCs and HS68 fibroblasts upon UV-B irradiation especially after high UV-B doses (Fig. 6b). P53 levels were elevated already at 6 hours after irradiation and showed a dose-dependent response with higher levels after 100 mJ/cm 2 . Whereas prolonged p53 stability and increased p21 levels were observed even 24 and 48 hours after irradiation with 100 mJ/cm 2 UV-B, 25 mJ/cm 2 UV-B resulted in similar p53 and p21 protein levels especially at later timepoints.

Discussion
While MSCs have shown beneficial effects regarding the regeneration of UV-induced skin damage, the effects of UV irradiation on MSCs themselves are largely unknown. Here, we elucidated the influence of UV-B irradiation on the survival and functional abilities of MSCs as well as mechanisms how MSCs deal with UV-B-induced DNA damage. We could show for the first time that human MSCs are relatively resistant to UV-B treatment and largely preserve their stem cells' characteristics.
MSCs have been shown to reduce UV-induced skin damage by secreting paracrine factors including keratinocyte growth factor (KGF), basic fibroblast growth factor-1 (FGF-1) and vascular endothelial growth factor (VEGF), thereby promoting collagen and fibronectin production of dermal fibroblasts 23,24,30 . Preclinical and early clinical studies have demonstrated beneficial effects of MSC-based therapies for psoriasis and atopic dermatitis 25,26 . In a phase I/IIa study, patients with atopic dermatitis were treated with MSCs leading to a 50% reduction of the Eczema Area and Severity Index (EASI) score in the majority of patients 25 . Based on these encouraging www.nature.com/scientificreports www.nature.com/scientificreports/ results, a phase III trial now evaluates the efficacy of MSCs for atopic dermatitis (NCT03269773). The favorable results of MSC therapies for autoimmune skin disorders were mainly attributed to MSCs' paracrine effects and their immunomodulatory impact, e.g. through prostaglandin E2 and transforming growth factor-β1 (TGF-β1) 31 .
As the stem cells' ability to aid the regeneration of UV-damaged skin requires intact migratory functions, the observed reduction of cellular velocity after high-dose UV-B irradiation may reduce their regenerative effects. The decreased cellular motility in MSCs is in contrast to melanoma cells which exhibit increased cellular motility after UV-B irradiation via autocrine interleukin-8 (IL-8) secretion 32 . However, only low UV-B doses up to 30 mJ/ cm 2 were used in this study.
Data about the influence of UV-B irradiation on the adhesion ability of MSCs has been lacking so far. Although general adhesion ability of MSCs was maintained after UV-B irradiation, heterogenous results for MSCS derived from different donors were observed in our study. MSCs isolated from old donors have been shown to be more susceptible to ROS leading to reduced integrin expression, impaired adhesion ability and reduced engraftment rates in a myocardial infarct model; however, the donor's age of our MSC preparations was quite homogeneous and ranged between 20 years and 32 years 33 . UV-B treatment has shown to inhibit intercellular adhesion molecule 1 (ICAM-1) induction by γ-interferon in keratinocytes at early time points, while it results in ICAM-1 induction at later time points beginning 48 hours after exposure 34 . A similar biphasic response was observed for melanocytes and melanoma cells: Cytokine-induced ICAM-1 synthesis was inhibited within the first 16 hours and increased between 48 and 96 hours after UV-B treatment 35 . The heterogeneous responsiveness of MSCs adhesion ability towards UV-B irradiation may be one possible explanation for different results upon therapeutic UV irradiation in autoimmune skin diseases.
Proliferation and cellular viability in vitro can partly predict the regenerative capacity of MSCs in vivo, and the increased proliferation and viability values compared to dermal fibroblasts are a promising indicator of an intact regenerative capacity 36 . Interestingly, metabolic viability determined by MTS assays was relatively well preserved after high UV-B doses up to 1500 mJ/cm 2 UV-B irradiation, while clonogenic survival rates were below 5% compared to untreated controls at 200 mJ/cm 2 . Obviously, cellular reproductive death occurs at much lower UV doses than impairment of metabolic viability. Clonogenic survival assays are commonly used for evaluation of radiation sensitivity and determine the cells' ability to undergo multiple cell divisions, whereas MTT and MTS assays are rather applied to study cellular chemosensitivity by measuring cell proliferation and intact mitochondrial respiration 37,38 . A similar discrepancy between clonogenic survival and metabolic viability of MSCs was observed for some chemotherapeutic agents such as paclitaxel and topoisomerase inhibitors 39,40 .
MSCs exert their regenerative effects via differentiation into functional cells and the creation of a protective microenvironment. Therefore, their preserved differentiation capacity is a further surrogate parameter for sufficient regenerative abilities of MSCs after UV irradiation. Our data suggest that endogenous dermal MSCs may preserve their regenerative abilities after single dose UV-B exposure. However, it is challenging to investigate whether multiple exposures to UV-B radiation induce different effects on MSCs, as they exhibit a short culturing time due to premature senescence in vitro 41 . www.nature.com/scientificreports www.nature.com/scientificreports/ Previous studies have demonstrated that MSCs exhibit varying sensitivities to different DNA-damaging agents; however, a relative resistance has been shown for the majority of DNA-targeting chemotherapeutical agents 39,[42][43][44][45] . Efficient DNA damage repair and high expression of anti-apoptotic proteins contribute to the cells' ability to evade apoptosis 46,47 . The observed low apoptosis levels after UV-B irradiation are consistent with previous reports showing that MSCs may evade apoptosis induction by undergoing premature senescence 48 . Accordingly, we detected significantly increased β-galactosidase expression in two MSC samples. Senescent MSCs exhibit a reduced regenerative potential with a compromised immunoregulatory capacity 49 . Whether UV-B-mediated senescence induction in MSCs may impair the regenerative ability of MSCs in vivo needs to be investigated in further studies.
Previous studies have examined the effects of varying UV-B doses on the stability, modifications and activity of the tumor suppressor p53 50 . P53 protein levels are generally low due to constant degradation via ubiquitin-dependent proteolysis 51 . While low UV-B doses generally result in fast but transient p53 accumulation, higher UV-B doses lead to delayed but prolonged p53 protein level increase 52 . However, both doses used in our study (25 mJ/cm 2 and 100 mJ/cm 2 ) are considerably lower than required for sustained p53 accumulation (e.g. 350 mJ/cm 2 UV-B were used in the study by Latonen et al. to induce a prolonged p53 increase 52 ). In line with this, p53 levels were comparable between untreated and UV-irradiated MSC3 cells and only slightly elevated in UV-irradiated MSC1 48 hours after UV exposure. P21 is a downstream effector of p53 and inhibits several cyclin-dependent kinases (CDKs), finally leading to cell cycle arrest 53 . Additionally, p21 has a crucial role in senescence induction and has been shown to protect from p53-mediated apoptosis 54,55 . Accordingly, p21 levels were considerably increased in MSCs 24 and 48 hours after high dose (100 mJ/cm 2 ) UV-B treatment which could at least partly explain increased senescence and low apoptosis rates of MSCs.
Nucleotide excision repair (NER) has been established as the main repair mechanism of UV-induced photoproducts such as CPDs and 6-4PPs 56 . Several readouts including T4 endonuclease V activity, UV-induced DNA repair synthesis or CPD repair measurements have been used to quantify NER after UV-B irradiation 57,58 . In our study, determination of NER activity was based on quantifying CPD levels after treatment. In accordance with previous reports demonstrating an effective NER activity of MSCs, we found that MSCs were able to repair UV-B-induced CPDs more efficiently than dermal fibroblasts 59 . In line with these findings, another study reported efficient DNA glycosylase activity in cultured adipose tissue-derived MSCs 60 .
In our dataset, human MSCs derived from bone marrow samples were used. It is conceivable that our results are also valid regarding skin-derived MSCs as both types have been demonstrated to show similar cellular morphology, surface markers expression and differentiation ability 61 . However, this hypothesis needs corroboration in further studies.
While several results were highly consistent after UV-B treatment for all tested MSC samples, including increased survival rates, impaired cellular velocity, stable surface marker expression, elevated osteogenic differentiation potential and low apoptosis levels, some results in our dataset revealed a considerable heterogeneity between individual MSCs. Regarding the increased senescence rates for MSC2 after UV-B exposure, we cannot rule out a potential impact of the donor's age on the induction of cellular senescence. Some studies have shown a link between donor's age and senescence levels in vitro, so that the increased age of donor #2 (32 years) compared to the other donors (20 years for donor#1, 25 years for donor#3) may contribute to the different senescence rate after UV-B treatment 62 . The known heterogeneity of MSCs derived from different donors may also need to be taken into account for the application of MSC-based treatments for skin diseases.
Prior to routine clinical application of MSC-based therapies for UV-induced skin damage or autoimmune skin diseases, any pro-tumorigenic potential especially for skin cancer must be thoroughly ruled out. It has been previously shown that growth of B16 melanoma cells was enhanced in the presence of co-injected MSCs, especially when MSCs were pre-incubated with interferon-γ and tumor necrosis factor-α which led to increased expression of the immunosuppressive enzyme inducible nitric oxide synthase (iNOS) in MSCs 63 . As iNOS inhibition abrogated the tumor-promoting effects of MSCs, the immunosuppressive abilities of MSCs may be one reason for the observed pro-tumorigenic effects of these cells in the model used. However, other studies have reported contrary effects of MSCs on melanoma cells such as reduction of proliferation in vitro and inhibition of tumor growth in vivo 64 .
Besides the known DNA-damaging potential, UV-B irradiation is also able to generate ROS, leading to oxidative damage and skin carcinogenesis as a potential long-term result 65 . Although we have not examined the anti-oxidative capacity of MSCs after UV-B treatment, several publications have reported the stem cells' ability to efficiently inactivate ROS due to high glutathione and superoxide dismutase levels 66 . The efficient antioxidative capacity of MSCs may contribute to the observed UV-B resistance of MSCs; however, further experiments are needed to elucidate the role of the antioxidative capacity in terms of the stem cells' UV-B response.
Ambient UV exposure comprises mainly UV-A irradiation at a wavelength of 315 to 400 nm which is, compared to UV-B irradiation, less intense but penetrates more deeply 67 . A limited number of studies examined the effects of UV-A irradiation on MSCs and revealed reduced adipogenic differentiation capacity but unchanged gene expression after UV-A-exposure 68,69 . UV-A acts mainly via indirect and ROS-mediated DNA damage, and the DNA-damaging effect of UV-A is less pronounced than that of UV-B. Therefore, our results may not be completely transferrable to the response of MSCs to UV-A irradiation.
In summary, our findings indicate a UV-B resistant phenotype of MSCs which may contribute to MSCs' ability to attenuate UV-B-induced skin injuries. Considering the UV-protective and immunomodulatory properties of MSCs, our data may warrant further analyses regarding combination studies of MSCs and UV-B irradiation for the treatment of autoimmune skin diseases.

Materials and Methods
Cell culture. Human MSCs were isolated from the bone marrow of three healthy donors as described before (MSC1: male donor (20 years old), MSC2: male donor (32 years old), MSC3: male donor (25 years old) 70 . Informed consent was obtained prior to bone marrow aspiration, and this investigation was approved by the www.nature.com/scientificreports www.nature.com/scientificreports/ Heidelberg University ethics committee (#S-384/2004). Human HS68 dermal fibroblasts were purchased from the ATCC (Manassas, USA). Cells were maintained at 37 °C in a humidified incubator with 5% CO 2 . Mesenchymal Stem Cell Growth Medium (Lonza, Basel, Switzerland) was used for culturing MSCs, while HS68 cells were grown in Dulbecco's Modified Eagle's Medium (Biochrom, Berlin, Germany) supplemented with 10% fetal bovine serum. All analyses of this study were performed in accordance with the relevant guidelines and regulations.

UV-B irradiation. UV-B irradiation was performed using a Waldmann UV181BL source (Waldmann,
Villingen-Schwenningen, Germany) with an output range of 280-320 nm wavelength. For each treatment, exact UV doses were measured with an UV detector (Waldmann). As the initial UV-B broadband dose used in psoriasis treatment is between 20 and 60 mJ/cm 2 , and minimal erythema dose normally ranges between 80 and 240 mJ/ cm 2 , a low-dose (25 mJ/cm 2 ) and a high-dose (100 mJ/cm 2 ) treatment group were used in the experiments 71,72 , except of assays for clonogenic survival, proliferation and viability where doses up to 1500 mJ/cm 2 where used.
Clonogenic, proliferation and viability assays. For clonogenic survival assays, between 400 and 1800 cells were plated in 6-well plates prior to treatment and allowed to grow for 14 days. Colonies were fixed with 25% acetic acid in methanol and stained with crystal violet solution. Colonies containing more than 50 cells were counted using an inverted Leica DM IL microscope (Leica Microsystems, Wetzlar, Germany), and the survival fraction was calculated as follows: (#colonies/#plated cells) treated /(#colonies/#plated cells) untreated . Experiments were performed with three biological triplicates.
To investigate the proliferation activity and viability after UV-B irradiation, between 3 × 10 4 and 4 × 10 4 cells were seeded in 6-well plates, and UV-B irradiation was performed 24 hours later. At 96 hours after irradiation, cells were harvested and stained with trypan blue to count viable cells using a Neubauer chamber.

Cell adhesion measurements.
Cells were grown in Petri dishes to a confluence of 70% prior to UV-B irradiation. Immediately after treatment, 100 cells/well were seeded in 96-well plates, and the number of attached cells was counted at different time points. The ratio between attached and seeded cells was calculated for each time point to determine the adhesion rate. Cellular velocity assays. Cellular velocity was measured by time-lapse microscopy. 2 × 10 4 were plated in Petri dishes (10 cm diameter) prior to UV-irradiation. Time-lapse microscopy was conducted on the Keyence BioRevo9000 microscope (Keyence, Neu-Isenburg, Germany) fitted with an incubator box at 37 °C and 5% CO 2 . Manual single-cell tracking with ImageJ software (National Institutes of Health, Bethesda, USA) was used for quantification, and at least 10 cells/well from three randomly chosen fields-of-view were tracked. Surface marker expression. Cells were plated in Petri dishes (diameter 10 cm) to a confluence of 70% and UV-irradiated. At 96 hours after UV-B irradiation, cells were harvested and fixed with 3% paraformaldehyde in PBS. MSC surface marker expression was analyzed using the MSC Phenotyping Kit (Miltenyi Biotec, Bergisch-Gladbach, Germany) following the manufacturer's instructions. Surface marker expression was determined on a FACSCanto TM flow cytometer (BD, Heidelberg, Germany), and data analysis was performed with FlowJo 7.6.5 software (FlowJo LLC, Ashland, USA).
For osteogenic differentiation, 25000 cells were plated on glass cover slips in 24-well plates and UV-irradiated 24 hours later. STEMPRO ® Osteogenesis differentiation medium (Gibco) was used to induce osteogenic differentiation. For quantification, specimens were incubated with OsteoImage ™ Staining Reagent (Lonza) according to the manufacturer's instructions.
Chondrogenic differentiation was performed using the STEMPRO ® Chondrogenesis Differentiation Kit (Gibco). After UV-B irradiation, 1 × 10 5 cells/well were plated in 96-well plates in order to induce spheroids. 21 days later, spheroids were fixed with 4% paraformaldehyde in PBS, frozen at −20 °C and sectioned on a cryomicrotome. Sections were incubated with 0.3% Triton X-100, 1% BSA and 10% normal donkey serum in PBS prior to incubation with an antibody against human aggrecan (1:10, R&D Systems, Minneapolis, MN, USA) and counterstaining with an Alexa488-coupled secondary antibody (1:200, Donkey Anti-Goat; Abcam, Cambridge, UK). For all differentiation experiments, fluorescence images were obtained from with a Keyence BioRevo9000 microscope, and staining intensities were normalized to cell numbers. www.nature.com/scientificreports www.nature.com/scientificreports/ apoptosis rate were assessed with a LSR II flow cytometer (BD), and data analysis was carried out using FlowJo 7.6.5 as reported before 73 . Experiments were performed with three replicate samples.

Cell cycle and apoptosis measurements.
Senescence analyses. 2 × 10 3 cells were seeded on each glass cover slip in a 24-well plate prior to UV-B treatment. At various time points after irradiation, cells were fixed, and β-galactosidase activity was measured using the Senescence β-galactosidase Staining Kit (Cell Signaling Technology, Leiden, Netherlands) following the manufacturer's instructions. Images were obtained with a Keyence BioRevo9000 microscope, and assessment of β-galactosidase-positive cells was performed using ImageJ. CPD ELISA assays. Cells were UV-B irradiated, and at different time points after treatment, DNA was isolated using the QIAamp ® DNA Mini Kit (Qiagen, Hilden, Germany). Cells were incubated with proteinase K (activity: 600 mAu/mL) and lysis buffer at 56 °C for 10 minutes, before 100% ethanol was added to the sample.
DNA was purified using QIAamp ® Mini spin columns, and DNA concentration was quantified by NanoDrop (Thermo Scientific, Wilmington, DE, USA). Analysis of CPD repair was performed using the High Sensitivity CPD ELISA Kit (Cosmo Bio Co., Tokyo, Japan) according to the manufacturers' protocol. Briefly, DNA was coated to the plate, and after washing and blocking of non-specific antibody binding, wells were incubated with anti-CPD antibody (1:100, clone TDM-2). Following incubation with the biotinylated secondary antibody (1:100), streptavidin-peroxidase was added. Wells were incubated with O-phenylenediamine, and absorbance at 492 nm was determined using a SPECTROstar Nano microplate reader (BMG LABTECH, Ortenberg, Germany).

Western blots.
MSCs and HS68 dermal fibroblasts were grown in Petri dishes to a confluence of 80% and then exposed to UV-B radiation. Cells were harvested at various time points after treatment, and cell pellets were incubated in RIPA buffer for 20 minutes on ice. Protein samples were run on 12% tris-acetate gels and transferred to polyvinylidenedifluoride membranes (Millipore, Darmstadt, Germany). After blotting, membranes were incubated with antibodies against p53 (1:1000, Cell Signaling) and p21 (1:500, BD Pharmingen), while β-actin (1:500, MP Biomedical, Solon, OH, USA) was used as a loading control. After incubation and several washing steps, membranes were incubated with the HRP-conjugated secondary antibodies anti-mouse-HRP (1:1000, W402B, Promega) and anti-rabbit-HRP (1:1000, W401B, Promega). Western blots were visualized on X-ray films using a Luminol-based enhanced chemiluminescence (ECL) HRP substrate (Thermo Scientific ™ SuperSignal ™ West Dura Chemiluminescent Substrate) following the manufacturer's instructions.
Statistics. At least three experimental replicates were carried out to calculate mean values and standard deviations. Comparisons between control and treatment group were performed using unpaired, two-sided Student's t-tests. P-values < 0.05 were considered significant.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.