Oct 30, 2024

Abstract Radiation-induced morphea (RIM) is a rare complication of radiotherapy presenting as inflammatory fibrosis, most commonly reported in breast cancer patients. As underlying disease mechanisms are not well understood, targeted therapies are lacking. Since fibroblasts are the key mediators of all fibroproliferative diseases, this study aimed to characterize patient-derived fibroblasts to identify therapeutic targets. We studied primary human control and RIM-fibroblasts on a functional and molecular basis, analyzed peripheral blood and tissue samples and conducted, based on our findings, a treatment attempt in one patient. In RIM, we identified a distinct myofibroblast phenotype reflected by increased alpha-smooth-muscle-actin (αSMA) expression, reduced proliferation and migration rates, and overexpression of osteopontin (OPN). Our RNA sequencing identified aberrant Myc activation as a potential disease driver in RIM fibroblasts, similar to previous findings in systemic sclerosis. Treatment with the anti-inflammatory drug mesalazine reversed the myofibroblast phenotype by targeting Myc. Based on these findings, a patient with RIM was successfully treated with mesalazine, resulting in reduced inflammation and pain and tissue softening, while serum OPN was halved. The present study provides a comprehensive characterization of RIM fibroblasts, suggests a disease-driving role for Myc, demonstrates promising antifibrotic effects of mesalazine and proposes OPN as a biomarker for RIM. Introduction Radiation-induced Morphea (RIM) is a rare and potentially under-recognized complication most commonly reported in breast cancer patients after completion of radiotherapy 1 . RIM is frequently misdiagnosed for cellulitis, mastitis or cancer recurrence 1 . Therefore, an early biopsy is important in distinguishing RIM from phenotypically similar diseases. RIM typically displays a superficial lymphocytic infiltrate and increased collagen deposition in the reticular dermis possibly accompanied by eosinophilia 2 , 3 . Recent epidemiological data suggests that 1 out of 378 breast cancer patients 4 is affected. The clinical appearance comprises erythema, edema and tissue induration as well as shrinkage of the affected breast, physical and emotional pain, ulceration and resulting infections 5 , 6 . Neither dose nor fractionation of radiotherapy seem to correlate with the incidence or severity of RIM 4 , 5 . Although an association to autoimmune disease has been proposed 6 , the knowledge about the underlying cellular and molecular disease mechanisms remains elusive as of today. Steroids, methotrexate, phototherapy and surgery can lead to varying symptomatic relief 1 . However, as found in systemic sclerosis treatment approaches, fibrosis is a generally poorly responding symptom 7 . Thus, a cellular and molecular characterization of RIM pathophysiology to derive novel treatment strategies is warranted. Fibroblasts have emerged in the spotlight of antifibrotic drug research, as they are key cellular regulators of the connective tissue. Fibroblasts maintain the balance between extracellular matrix (ECM) synthesis and degradation thereby contributing to ECM and tissue homeostasis 8 , 9 , 10 . Upon exposure to activating stimuli like radiation 11 , chemical mediators or mechanic tissue injury 12 , fibroblasts undergo a phenotypic transition towards myofibroblasts, characterized by the expression of organized α-smooth muscle actin (αSMA) microfilaments 8 , 9 . Myofibroblasts promote and maintain tissue fibrosis by secretion of vast amounts of ECM and a multitude of inflammatory mediators like transforming growth factor beta (TGFβ) and interleukins 13 , 14 . Currently, there are numerous therapeutic efforts aiming at preventing or even reversing fibrosis, including antibody therapies, small molecules and cellular immunotherapies 15 , 16 , 17 , 18 . However, in order to offer new antifibrotic therapies in a timely manner, the use of an already existing, FDA or EMA approved drug with a favorable side effect profile would be optimal (drug repurposing). First evidence has indicated that systemically administered mesalazine (5-ASA), which is clinically used to treat chronic inflammatory bowel disease, is capable of stopping and partly reversing fibrosis progression in a model of induced liver fibrosis 19 . Furthermore, we recently demonstrated that mesalazine is sufficient to prevent cardiac and dermal fibrosis in vitro and in mouse models 20 , 21 , 22 , 23 . Based on the functional and molecular characterization of RIM-derived fibroblasts, we identified a potentially disease-driving role for aberrant Myc activation in RIM fibroblasts. Moreover, we provide experimental as well as clinical evidence for repurposing of mesalazine as a novel treatment approach. Results Characterization of control and RIM-derived cultured fibroblasts Primary skin fibroblasts were isolated via outgrowth 10 from skin biopsies of 5 control and 5 RIM patients and subsequently seeded on glass coverslips for cytomorphological characterization. All cells displayed typical fibroblast morphology, reflected by a spindle shape and stellate processes (Figure S1 A). Accepted fibroblast marker proteins like vimentin, human fibroblast surface protein (hFSP), Discoidin domain-containing receptor 2 (DDR2) and collagen1 9,10,24 were present in immunocytochemical analysis (Figure S1 b-e). Myofibroblast differentiation is a critical step in fibrosis development 8 . Compared to control, basal expression of the myofibroblast marker αSMA as determined by immunostaining was 16.1% higher in RIM (αSMAcontrol: 7.3 ± 0.68%; αSMARIM: 23.4 ± 1.81%) (Fig. 1 a). This finding was validated by western blot, which demonstrated significantly higher (p = 0.0132) αSMA protein expression in RIM (Fig. 1 b). Functionally, RIM-derived fibroblasts displayed markedly reduced proliferation rates and reduced migration in wound healing assay compared to control indicating their reduced wound healing potential (Fig. 1 c and d). Senescence-associated beta-galactosidase activity 25 was not detectable in either group. Lastly, exemplary immunohistochemistry revealed diffuse tissue αSMA expression in a RIM-derived skin punch biopsy suggesting exaggerated myofibroblast presence, whereas an αSMA signal was only detectable around blood vessels in the respective control sample (Fig. 1 e). Fig. 1 Radiation-induced morphea is accompanied by a pronounced myofibroblast phenotype. (a) Quantification and representative images of αSMA (alpha smooth muscle actin; red) immunofluorescence staining in primary patient-derived control and RIM (radiation induced morphea) fibroblast cultured under basal conditions (37 °C, 5% CO2, Dulbecco’s modified eagle medium supplemented with 10% fetal calf serum and 1% penicillin-streptomycin) for 7 d, the nuclei were stained with DAPI (blue), (n = 10 analyzed coverslips from N = 5 patients per group; results are given as mean ± SEM determined by a Mann-Whitney U test). The scale bars equal 200 μm. (b) Quantification and representative western blot of αSMA protein abundance in control and RIM (radiation induced morphea) fibroblast, normalized to GAPDH, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test). (c) Proliferation curves of control and RIM-fibroblasts under basal conditions, (n = 8 from N = 4 patients per group, results are given as mean ± SEM determined by a Welch t test for each time point). (d) Number of migrated control and RIM-fibroblasts in a wound healing assay after 24 h, (n = 8 from N = 4 patients per group, results are given as mean ± SEM determined by a Mann-Whitney U test). (e) Exemplary immunofluorescence images of αSMA (red) in skin punch biopsies of a control and a RIM patient; the nuclei were stained with DAPI (blue). The scale bars equal 100 μm. Gene expression analysis We performed mRNA sequencing (RNA-Seq) of primary control and RIM-derived patient skin fibroblasts to explore putative differences in gene expression with particular emphasis on myofibroblast differentiation, cytokines and extracellular matrix as well as regulatory pathways. Differential gene expression analyses revealed a total of 3641 differentially expressed genes (padjusted < 0.05), that split into 1792 up-regulated and 1849 down-regulated genes when comparing the RIM samples to the control group (Fig. 2 a-c; Table S1 ). Gene sets focusing on fibrosis mechanisms for subsequent analysis were provided by GSEA 26 , 27 , 28 and The Harmonizome 29 . Additionally, we performed pathway analysis with Metascape to uncover the functional relevance of differentially expressed genes 30 . In line with the results of the phenotypic characterization, RIM-fibroblasts showed a significantly (p = 0.0095) higher expression of ACTA2 (αSMA). We found robust expression levels of matrix metalloproteinase (MMP) 1, 2, 3 and 14 with significantly lower MMP2 (p = 9.357e-05) and MMP14 (p = 0.0017) expression in RIM compared to control (Fig. 2 a, b; Table S1 ). There was no difference in the expression of transforming growth factor beta (TGFβ, TGFB1) (p= 0.1518), which is widely regarded as the “master regulator” of fibrosis 31 . However, the TGFβ pathway showed significant enrichment (Fig. 2 e). In line with this finding, RIM-fibroblasts were found to have a significantly higher expression of the TGFβ downstream target osteopontin (OPN; Table S1 ), which is both a constituent of the extracellular matrix and a cytokine 32 , 33 relevant in cardiac, pulmonary, hepatic and dermal fibrosis 19 , 33 , 34 , 35 . Among known pro-fibrotic transcription factors, there was a significantly higher expression of Myc (p = 0.0028) and β-Catenin (CTNNB1, p = 7.4567e-06) in RIM (Table S1 ), consistent with previous findings in systemic sclerosis 36 , 37 . Enrichment Analysis of transcriptional regulatory networks confirmed significant enrichment of genes regulated by Myc (Fig. 2 d). Further disease association analysis with Metascape using DisGeNET 38 revealed enrichment of several cancer-associated pathways, in particular estrogen receptor positive breast cancer (Fig. 2 f). Lastly, we used the STRING database 39 , which integrates protein-protein interactions to uncover functional and regulatory interactions, to create a concise interaction network of identified genes of interest (Myc, CTNNB1, ACTA2, SPP1 (OPN)), which have been associated with fibrosis indicating indirect profibrotic roles for Myc and CTNNB1(Fig. 2 g). Fig. 2 Differential gene expression analyses and FPKM value based clustered heatmaps comparing RIM samples with the control group. a) Deseq2 results depicted in a volcano plot. Significantly up-regulated genes are depicted in red, significantly down-regulated genes are depicted in blue. Genes which are listed in the myofibroblast differentiation (b) dataset are highlighted with a black cross. The x-axis shows the log2FoldChange and the y-axis the negative log10 value of the adjusted p-value for each gene. b, c) Expression value based clustered heatmaps of three manually selected gene sets. Clustering of RIM and control samples shows a clear separation between the two groups. Comprehensive gene lists with expression values for each gene set of interest are provided in Table S1 . d) Metascape analysis of significantly enriched transcriptional regulatory interactions. e) Network of enriched terms using Cytoscape5 78. Each node represents an enriched term and is colored by its cluster ID. f) Metascape analysis of significantly enriched disease associated pathways. g) Protein-protein interaction network with the fibrosis-associated genes Myc, CTNNB1, ACTA2, SPP1 based on textmining, experimental data, databases and co-expression with high confidence (0.700) using the STRING database. The line thickness indicates the strength of data support. OPN function in RIM Based on the RNA-seq data, we went on to characterize the function of OPN in RIM, as exaggerated OPN expression has been linked to systemic sclerosis and other fibrotic diseases 33 , 35 . Both OPN mRNA and protein expression were significantly higher in RIM compared to control fibroblasts (p = 0.0310 and p = 0.0079, Fig. 3 a and b). In order to investigate whether OPN acts as either a disease driver or a potential biomarker in RIM, we exposed dermal HFF1 fibroblasts to recombinant OPN in ascending concentrations and performed an OPN knockdown in HFF1 and RIM fibroblasts in parallel. Although, we found a concentration-dependent increase in fibrillary αSMA expression in reaction to recombinant OPN, (Fig. 3 c), the knockdown of the endogenous OPN expression (Fig. 3 d) did not affect the cellular myofibroblasts phenotype. In HFF1, mRNA expression of ACTA2, COL1A1 and FAP (fibroblast activation protein alpha) remained unaltered (Fig. 3 e). In RIM fibroblasts, protein expression of αSMA and Collagen 1 showed similar results (Fig. 3 f), indicating that the profibrotic phenotype is not revertible by downregulation of OPN. Fig. 3 Characterization of OPN function in RIM. (a) Expression of OPN mRNA (qPCR) normalized to EEF2 (Eukaryotic translation elongation factor 2) as housekeeping gene in control and RIM-fibroblasts, (n = 5 per group; results are given as mean ± SEM determined by a Mann-Whitney U test). (b) Quantification and representative western blot and of OPN protein abundance in control and RIM fibroblast, normalized to EEF2 as housekeeping protein, (n = 5 per group; results are given as mean ± SEM determined by a Mann-Whitney U test). (c) Left: Quantification of αSMA staining. HFF1 fibroblasts were kept at the above indicated concentrations of recombinant OPN (Sigma-Aldrich, SRP3131) for 72 h. Results are given as mean ± SEM determined by One-Way ANOVA with Tukey post test. Right: Representative immunofluorescence images for αSMA of OPN-treated HFF1 fibroblasts. The nuclei were stained blue (DAPI). The scale bar equals 20 μm. (d) Expression of OPN mRNA (qPCR) normalized to HPRT (Hypoxanthine-guanine phosphoribosyltransferase) as housekeeping gene in HFF1 fibroblasts under control conditions and after OPN-knockdown with 3 different siRNAs, (n = 3 per group; results are given as mean ± SEM). (e) Expression of ACTA2 (αSMA), COL1A1 (collagen 1) and FAP (fibroblast activation protein alpha) mRNA (qPCR) normalized to HPRT in HFF1 fibroblasts 24 h after OPN knockdown (n = 3 per group; results are given as mean ± SEM determined by a Mann-Whitney U test for each gene). (f) Quantification and representative western blots for αSMA and Collagen 1 protein abundance in RIM fibroblasts ± OPN knockdown, normalized to GAPDH, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test). Characterization of the regulatory effects of β-catenin and myc signaling on the myofibroblasts phenotype in RIM Since aberrant β-Catenin and Myc signaling have been shown to exert profibrotic effects in systemic sclerosis, pulmonary and renal fibrosis and induce OPN expression 37 , 40 , 41 , 42 , 43 , 44 , 45 , we performed qPCR to confirm the RNA-Seq results and found that β-Catenin and Myc mRNA were significantly higher expressed in RIM compared to control (p = 0.0462 and p = 0.0414; Fig. 4 a and b). Further histological analysis on available samples from two RIM patients demonstrated expression of αSMA, β-Catenin, Myc and OPN (Figure S2 ). To establish a link between irradiation and the identified fibroblast phenotype in RIM, control fibroblasts were exposed to a single dose of 4 Gy X-irradiation and subsequently analyzed after 24 h. As determined by qPCR, we found a distinct increase in CTNNB1, MYC, ACTA2 and OPN mRNA expression induced by X-irradiation (Fig. 4 c). Since, the biological activity of β-Catenin and Myc is linked to their respective phosphorylation status 46 , we performed phosphorylation analysis by western blot. With regard to β-Catenin phosphorylation, no significant differences were detectable between control and RIM fibroblasts (Fig. 4 d). For Myc activity, two phosphorylation sites are considered particularly important. Phosphorylation at serine 62 has been linked to increased Myc activity, whereas phosphorylation at threonine 58 leads to Myc degradation and thereby to decreased activity 46 . We found a 3.7-fold higher Myc phosphorylation at serine 62 in RIM fibroblasts compared to control (p = 0.0126), while there was no significant difference in threonine 58 phosphorylation (Fig. 4 e). Therefore, subsequent experiments focused on the effects of pharmacological Myc modulation on the myofibroblast phenotype in RIM. First, we sought to answer the question whether pharmacological activation of Myc signaling (BML-284 47; 1, 3 and 10 µM for 8 h) was sufficient to induce a myofibroblast phenotype in HFF1 fibroblasts. Indeed, both αSMA and OPN protein expression increased in a concentration-dependent manner after 8 h of pharmacological treatment with BML-284 (Fig. 4 f). Fig. 4 Characterization of the regulatory effects of β-Catenin and Myc signaling on the myofibroblasts phenotype in RIM. (a) Expression of CTNNB1 (β-Catenin) mRNA (qPCR) normalized to EEF2 as housekeeping gene in control and RIM-fibroblasts, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test). (b) Expression of Myc mRNA (qPCR) normalized to EEF2 in control and RIM-fibroblasts, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test). (c) Radiation-dependent expression of CTNNB1, Myc, ACTA2 (αSMA) and OPN mRNA (qPCR) normalized to EEF2 in control fibroblasts after 24 h of exposure to 4 Gy of gamma irradiation, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test for each gene). (d) Quantification and representative western blots of β-Catenin phosphorylation at Ser552 and Thr41/Ser45 in control and RIM fibroblast normalized to GAPDH, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test). (e) Quantification and representative western blots of Myc phosphorylation at Ser62 and Thr58 in control and RIM fibroblast normalized to GAPDH, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test). (f) Quantification of αSMA protein abundance in HHF1 fibroblasts upon pharmacological activation of Myc signaling with 1, 3 and 10 µM BML-284 or vehicle control (1 µL DMSO/mL medium), normalized to GAPDH, (ncontrol = 4, nBML-284 = 3 per group; results are given as mean ± SEM determined by a One Way ANOVA and Dunnett’s multiple comparison test). (g) Quantification of OPN protein abundance in HHF1 fibroblasts upon pharmacological activation of Myc signaling with 1, 3 and 10 µM BML-284 or vehicle control (1 µL DMSO/mL medium), normalized to GAPDH, (n = 6 per group; results are given as mean ± SEM determined by a One Way ANOVA and Dunnett’s multiple comparison test). (h) Representative western blots for f and g. Mesalazine modulates the cellular localization and phosphorylation status of myc As we found increased Myc activity in RIM fibroblasts and pharmacological experiments indicated a positive correlation of Myc activation and the induction of myofibroblast differentiation (Fig. 4 f-h), our aim was the modulation of this signaling pathway with an already clinically approved compound. Mesalazine is used for the treatment of inflammatory bowel disease and has recently been shown to exert antifibrotic effects in vitro and in vivo 22 , 23 , 24 , 19 . The exact mechanism of action cannot be attributed to the modulation of a single pathway 20 . Inhibition of NFκB, activation of PPARγ, inhibition of TGFβ-SMAD-signaling but also inhibition of Myc signaling have been reported 24 , 48 . First, we examined the effects of mesalazine on the cellular localization of Myc. Immunofluorescence analysis of HFF1 fibroblasts showed that Myc was mainly located in the cytoplasm, whereas pharmacological treatment with BML-284 resulted in an overall increased Myc fluorescence signal intensity and significantly increased nuclear localization (p = 0.0007; Fig. 5 a left and middle image). Notably, co-treatment with mesalazine reduced overall Myc fluorescence intensity and abolished the nuclear Myc signal (Fig. 5 a middle and left image, b). Probing for a potential mechanism behind these findings, we examined the Myc phosphorylation status after mesalazine treatment in RIM fibroblasts. After 24 h of mesalazine treatment, Ser62 phosphorylation remained unchanged but there was a significant increase in Thr58 phosphorylation (Fig. 5 c), which has been linked to Myc degradation 46 . In order to establish a causal relation between radiation and Myc localization, we exposed HFF1 fibroblasts to a single dose of 4 Gy X-irradiation ± mesalazine treatment and analyzed nuclear Myc expression after 24 h via immunofluorescence (Fig. 5 d). Radiation significantly (p < 0.0001) increased the nuclear Myc signal compared to non-irradiated HFF1 control fibroblasts. In line with the previous results, mesalazine treatment led to a significant (p = 0.049) reduction of the nuclear Myc signal. Fig. 5 Impact of mesalazine treatment on the cellular localization and phosphorylation status of Myc. (a) Representative immunofluorescence images of Myc (red) in human skin fibroblasts (HFF1) upon treatment with vehicle control (1 µL DMSO/mL medium); 1 µM BML-284 or 10 mM mesalazine subsequent to 1 µM BML-284. The nuclei were stained with DAPI (blue). The scale bars equal 20 μm. (b) Quantification of whole cell, cytoplasmic and nuclear Myc fluorescence intensity relative to the total cellular area determined using the CellProfiler software (version 4.2.1); (Whole cell [n = 66, 108, 62], Cytoplasma [n = 65, 110, 62], Nucleus [n = 67, 114, 63], results are given as mean ± SEM determined by a Brown-Forsythe and Welch ANOVA test). (c) Quantification and representative western blots of Myc phosphorylation at Ser62 and Thr58 normalized to GAPDH in RIM fibroblast under basal conditions and after 10 mM mesalazine treatment for 24 h, (n = 5 per group; results are given as mean ± SEM determined by a Welch t test). (d) Quantification and representative images of nuclear Myc fluorescence intensity relative to the total cellular area in HFF1 fibroblasts under control conditions, 4 Gy x-irradiation ± 10 mM mesalazine treatment for 24 h [n = 103, 251, 141]. The nuclei were stained with DAPI (blue). The scale bars equal 20 μm. Results are given as mean ± SEM determined by a Brown-Forsythe and Welch ANOVA test). Mesalazine reverses the RIM phenotypein vitroand in a patient We went on to evaluate the mesalazine effects on the myofibroblast phenotype in vitro.As antifibrotic effects of mesalazine have been described previously 19 , 22 , 23 , 24 , we investigated αSMA and OPN protein expression by western blot in primary control, RIM and mesalazine-treated RIM-fibroblasts. After 72 h, no significant difference in αSMA protein expression was found between control and mesalazine-treated RIM-fibroblasts, whereas αSMA expression was significantly elevated in untreated RIM-fibroblasts compared to control (p = 0.0033; Fig. 6 a left panel). A similar effect was observed for OPN protein abundance (Fig. 6 a middle panel). Concomitantly, the exaggerated expression of fibrillary αSMA bundles in RIM-fibroblasts, as determined by immunocytochemical staining, was reduced to control levels after 72 h mesalazine treatment (Fig. 6 b). Additionally, we performed a western blot for collagen 1 protein expression in RIM fibroblasts at baseline conditions and after mesalazine treatment. Collagen 1 expression was significantly reduced in mesalazine-treated RIM fibroblasts as compared to untreated controls (p = 0.0268, Fig. 6 c). Fig. 6 Mesalazine treatment improves the cellular and clinical disease phenotype of RIM. (a) Quantification of αSMA (left panel) and OPN (middle panel) protein abundance normalized to EEF2 in primary patient-derived control and RIM fibroblast under basal conditions and after 10 mM mesalazine treatment for 72 h, (control [n = 5], RIM [n = 4], RIM + Mesa [n = 4]; results are given as mean ± SEM determined by a Kruskal-Wallis test with Dunn’s multiple comparisons test). Representative western blots are depicted in the right panel. (b) Quantification and immunofluorescence images of αSMA (red) in primary patient-derived control and RIM fibroblast under basal conditions and after 10 mM mesalazine treatment for 72 h, the nuclei were stained with DAPI (blue), the scale bars equal 50 μm, (control [n = 6 from N = 3 patients], RIM [n = 10 from N = 5 patients], RIM + Mesa [n = 10 from N = 5 patients]; results are given as mean ± SEM determined by a One Way ANOVA and Dunnett’s multiple comparison test). (c) Quantification and representative western blots of Collagen 1 normalized to EEF2 in RIM fibroblast under basal conditions and after 10 mM mesalazine treatment for 72 h, (n = 4 per group; results are given as mean ± SEM determined by a Welch t test).d) Graphic representation of the patient’s subjective pain level using the numeric rating scale. e) Representative images of the patient before and after oral mesalazine treatment (1 g twice daily) for 6 weeks. Left image: Retraction of the right breast with inflammatory erythema, edema, tissue atrophy and marked shiny lesions. The tissue was hardened upon palpatory examination. Right image: Erythema and edema were markedly reduced. The tissue was distinctly softer upon palpatory examination and shiny lesions were reduced. f) ELISA measurement of OPN concentration in the peripheral blood of healthy control patients and patients with RIM. One RIM patient received oral mesalazine treatment (1 g twice daily) for 6 weeks, (control [n = 12], RIM [n = 4], RIM + Mesa [n = 1]; results are given as mean ± SEM determined by a Mann-Whitney U test). Encouraged by these findings, we conducted an individual treatment attempt in one of the patients whose fibroblasts were characterized in the present study, as previous therapies had failed. We present the case of a 58-year-old Caucasian woman (BMI 34.9 kg/m2, non-smoker), who suffered from infiltrating ductal carcinoma of the right breast two years prior to RIM onset. After completion of chemotherapy (paclitaxel) and radiation therapy for 3 months, the patient was treated with exemestane. A complete list of the patient’s medication can be found in Table  1 . 5 months after completion of radiation therapy, she first noticed erythema, severe pain, edema, tissue induration and a rapid shrinking of the right breast. The changes in skin condition were confined to the radiation port. The diagnosis of RIM was made based on the clinical presentation and histological analysis. Initial treatment consisted of topical and systemic steroids, methotrexate, physical therapy and UVA1 treatment without significant clinical response. The patient was informed about the off-label use of mesalazine and gave written informed consent to be treated with orally administered mesalazine (1000 mg, twice daily) for 6 weeks. Telephone interviews to determine pain and tolerability were conducted at weeks 2 and 4. After 6 weeks, a clinical examination was performed. Systemic mesalazine treatment led to a striking improvement in pain (reduction from 8 to 2 NRS pain, Fig. 6 d), tissue softening upon palpatory examination and pronounced reduction of erythema and edema (Fig. 6 e). We applied a modified version of the LoSCAT (Localized Scleroderma Cutaneous Assessment Tool) 49 , 50 , 51 to quantify the skin changes during treatment in the single affected region. The modified LoSCAT can be found in Table S2 . Briefly, the patient achieved an 80% reduction of disease activity and a modest reduction of disease damage during the mesalazine treatment. No serious adverse events were reported and laboratory results indicated good tolerability. OPN plasma levels were determined by ELISA in blood samples that were obtained for routine monitoring of liver and kidney function as well as blood count. Compared to healthy controls (24.23 ± 2.3 ng/mL), RIM patients had significantly elevated plasma OPN (p = 0.0077; 40.61 ± 3.5 ng/mL). After 6 weeks of mesalazine treatment, plasma OPN was reduced to average control levels in our patient (Fig. 6 f). Medication was stopped later by the patient because of stable disease. Table 1 Medication of the study patient. Histology and image analysis. For histological analysis, paraffin embedded skin Sect. (5 μm) from representative control and RIM patients were generated by the department of pathology at Dresden University Hospital in the course of clinical diagnostics. A detailed description of deparaffinization and subsequent immunohistochemical staining was published recently 34 . Fluorescence images were acquired with a Keyence BZ-X710 All-in-One Fluorescence Microscope (Keyence Corporation of America, Itasca, USA). Osteopontin knockdown. Lipofectamine (Thermo Fisher, Waltham, Massachusetts, USA; #13778030) was used for siRNA transfection according to the manufacturer’s instructions. The OPN knockdown was performed using the siRNAs listed in Table  3 (adapted from Patent No. : EP 2 290 063 A1; Table  3 ): Table 3 OPN siRNAs.