Selective inhibitors of bromodomain BD1 and BD2 of BET proteins modulate radiation‐induced profibrotic fibroblast responses

Radiotherapy can induce various adverse effects including fibrosis in cancer patients. Radiation‐induced aberrant expression of profibrotic genes has been associated with dysregulated epigenetic mechanisms. Pan‐BET (bromodomain and extraterminal domain) inhibitors, such as JQ1 and I‐BET151, have been reported to attenuate the profibrotic response after irradiation. Despite their profound preclinical efficacy, the clinical utility of pan‐inhibitors is limited due to observed cytotoxicicities. Recently, inhibitors were developed that selectively target the first (BD1) and second (BD2) bromodomain of the BET proteins (iBET‐BD1 [GSK778] and iBET‐BD2 [GSK046]). Here, their potential to attenuate radiation‐induced fibroblast activation with low‐toxicity was investigated. Our results indicated that cell proliferation and cell cycle progression in fibroblasts from BJ cells and six donors were reduced when treated with I‐BET151 and iBET‐BD1, but not with iBET‐BD2. After irradiation, induction of DGKA and profibrotic markers, especially COL1A1 and ACTA2, was attenuated with all BET inhibitors. H3K27ac enrichment was similar at the DGKA enhancer region after I‐BET151 treatment and irradiation, but was reduced at the COL1A1 transcription start site and the ACTA2 enhancer site. iBET‐BD2 did not change H3K27ac levels in these regions. BRD4 occupancy at these regions was not altered by any of the compounds. Cell migration activity was measured as a characteristic independent of extracellular matrix production and was unchanged in fibroblasts after irradiation and BET inhibitor‐treatment. In conclusion, iBET‐BD2 efficiently suppressed radiation‐induced expression of DGKA and profibrotic markers without showing cytotoxicity. Thus BD2‐selective targeting is a promising new therapeutic avenue for further investigations to prevent or attenuate radiotherapy‐induced fibrosis.

Thus BD2-selective targeting is a promising new therapeutic avenue for further investigations to prevent or attenuate radiotherapy-induced fibrosis.
Radiation therapy for cancer can cause fibrosis by disrupting epigenetic control mechanisms that activate pro-fibrotic genes. Inhibitors that broadly target the bromodomain and extra-terminal (BET) domain family can combat this activity, but aren't clinically useful due to high toxicity. Here, the authors tested bromodomain-selective inhibitors and found that iBET-BD2, which targets the second bromodomain of BET proteins, lessens the activation of pro-fibrotic genes with only minor cytotoxicity. This inhibitor could be a promising option for reducing fibrosis in cancer survivors.

| INTRODUCTION
Radiation-induced fibrosis is a common late side-effect of radiotherapy in cancer patients. 1 Around 20% of breast cancer patients and 68% of head and neck cancer survivors suffer from subcutaneous fibrosis after radiotherapy. 2 Differential gene expression and epigenetic predisposition have been observed in radiation-induced fibrosis of lung and skin. 4,5 A recent study showed that differential DNA methylation at an enhancer of diacylglycerol kinase alpha (DGKA) was associated with fibrosis risk and modulated profibrotic gene expression after irradiation through an EGR1-DGKA-COL1A1/3A1 axis in human fibroblasts. 5,6 The profibrotic response was suppressed by targeting the histone acetyltransferase CBP/p300 or acetylation-sensitive bromodomain-and extraterminal domain (BET) proteins. 6 BET-family proteins such as bromodomain containing (BRD) 2, BRD3 and BRD4 contain two tandem bromodomains (BD1 and BD2) that allow binding to acetylated histones and act as epigenetic "readers" to facilitate transcription. Alteration of BET protein activity results in transcription reprogramming in cancer, in autoimmune, cardiovascular and metabolic diseases, and in embryonic development. 7,8 First-generation BET inhibitors such as I-BET151 target both bromodomains and are found to be anti-inflammatory and anti-cancer agents in preclinical models and clinical trials. 7 Their clinical usage is, however, hampered due to cytotoxicity, 9 probably due to interference with basic cellular functions such as cell growth and cell cycle progression occurring in all dividing cells. Bromodomain-selective BET inhibitors have been recently developed and experimental data suggest that individual bromodomain sites have different biological functions. 10 Here, we investigated whether the pan-BET inhibitor I-BET151 and the bromodomain-selective inhibitors iBET-BD1 and iBET-BD2 can target the EGR1-DGKA-COL1A1/3A1 axis and alleviate the radiationinduced production of profibrotic proteins in fibroblasts. Our findings may provide insights for the development of novel therapeutic opportunities to prevent or attenuate radiotherapy-induced fibrosis.

| Cell lines and culture
Human foreskin-derived BJ fibroblasts (CRL-2522, RRID:CVCL_3653) obtained from American Type Culture Collection and normal human dermal fibroblasts (NHDFs, L1, L3, L4 and H1-3) were cultivated as described. 6 NHDFs were established from female donors aged 47 to 84 years (median age, 66 years) at the Universitätsmedizin Mannheim, Germany, as part of the EURATOM/ESTRO GENEPI project. 11 Fibroblasts were outgrown from skin biopsies taken from the un-irradiated inner upper arm of donors. The methylation status of NHDFs was as described. 6 All cell lines were recently authenticated using short tandem repeat or SNP profiling by Multiplexion (Heidelberg, Germany) (December 2020), and all experiments were performed with mycoplasma-free cells.
Maximal DMSO concentration in assays was 0.1%.

| Flow cytometry analysis of cell cycle
Cells were collected and fixed in ice-cold 75% ethanol, permeabilized using 1% Triton X-100, treated with RNAse A (#19101, Qiagen, Hilden, Germany) and finally incubated with 25 μg/mL propidium iodide, (#P4864, SigmaAldrich). Samples were run on a BD FACSCanto II flow cytometer (BD Bioscience, San Jose, CA), and at least 10 000 events/sample were acquired. The cell cycle distribution was evaluated using FlowJo software version 10 (BD Bioscience).

| Cell proliferation and viability assay
CellTiter Blue reagent (#G8081, Promega, Madison, WI) was used to measure cell viability and reflected proliferation. Cells were cultivated in 96-well plates, treated as described, and further processed according to the manufacturer's protocol.

| RNA isolation and real-time quantitative reverse transcription PCR (RT-qPCR)
RNA isolation and quantitative real-time PCR (RT-qPCR) were performed as previously described 6 using universal probe library hydrolysis probes (Roche, Basel, Switzerland) on a LightCycler 480 system (Roche). Primer sequences are listed in Table S1. Gene expression was normalized to the average of two housekeeping genes, HPRT1 and GAPDH.

| Enzyme-linked immunosorbent assay (ELISA)
The secreted collagen 1a1 was quantified using the COL1A1 ELISA kit  2.7 | Antibody-guided chromatin tagmentation and real-time PCR (ACT-qPCR) The experiment was performed as described. 6 (Table S2) on a LightCycler 480.
Signals reflect the relative H3K27ac enrichment of treated samples vs nonirradiated, DMSO-treated control samples.

| Chromatin immunoprecipitation and realtime PCR (ChIP-qPCR)
The experiment was performed as described, 13 with some modifications. Briefly, nuclei were isolated, lysed, and sheared by using a Covaris M220 sonicator (Covaris, Brighton, UK). Antibody against BRD4 (#13440, Cell Signaling, Danvers, MA) and rabbit IgG (#pp64, Millipore) were used. Specific regions were quantified for enrichment as described for ACT-qPCR. Enrichment was expressed relative to the sample without IP (% of input).

| Cell migration assay
100,000 BJ cells in 2 mL medium per well were cultivated in 6-well plates and treated with 1 μM BET inhibitors or DMSO for 48 hours.
The cell monolayer was scratched using a 200 μL plastic pipette tip, irradiated with 0 or 6 Gy, and incubated for another 20 hours with fresh inhibitor-containing media. Cells were fixed using 70% ethanol and stained using 1% crystal violet. The wound area was photographed using an inverted microscope (Leica DM IL LED) and images were analyzed using ImageJ software (version 1.53c, Rayne Rasband, National Institutes of Health; https://imagej.net/software/imagej/). 14

| Statistics
Statistical significances were determined by two-tailed Student's ttest, and results with P-values <.05 were considered statistically significant. Data processing and statistical analyses were performed in R 4.0.5, and visualized using ggplot2 (3.3.4). 15 The graphical abstract was created with BioRender.com.

| Cell proliferation and cell cycle in BJ fibroblasts treated with BET inhibitors and ionizing irradiation
To investigate the effects of BET inhibitors in normal human fibro-

| Radiation-induced fibroblast activation is reduced in BET inhibitor-treated BJ fibroblasts
As the expression of profibrotic extracellular matrix proteins can be indicative of fibroblast transactivation, 16    iBET-BD2, with or without irradiation ( Figure 4D). We further observed that H3K27ac enrichment at the COL1A1 transcription start site (COL1A1_TSS) was reduced in I-BET151 treated BJ fibroblasts without irradiation, and this reduction was also detectable after irradiation ( Figure 4E). At the ACTA2 enhancer site (ACTA2_enh), enriched

| Histone modifications after inhibitor treatment in irradiated BJ fibroblasts
H3K27ac was only suppressed in I-BET151-treated cells after irradiation. BRD4 binding was unaltered with or without irradiation neither at the COL1A1_TSS or the ACTA2_enh regions ( Figure 4F). Taken together, I-BET151 showed stronger effects than iBET-BD2 on the H3K27ac pattern of our regions of interest. Of note, none of the drugs interfered with the BRD4 enrichment at these regions.

| Cell migration after BET inhibitor treatment and ionizing irradiation of BJ fibroblasts
An early step in wound healing requires that fibroblasts migrate into the wounded area to stabilize the tissue by excreting extracellular matrix proteins. 17  fibrosis, 19 in radiation-induced lung fibrosis, 20 in cardiomyocytes [21][22][23] and in renal interstitial fibroblasts. 24 BRD4 silencing or treatment with the pan-BET inhibitor JQ1 also impeded fibroblast activation in human dermal fibroblasts with low or high DNA methylation at the DGKA enhancer after bleomycin treatment or irradiation. 6,13 In our studies, the BD2-selective inhibitor iBET-BD2 showed potent inhibition of profibrotic markers induced by irradiation with no detectable effects on cell proliferation and cell cycle progression in the investigated fibroblasts, pointing at a low cytotoxicity of this compound which might be beneficial for patient treatment.
Important hints for toxic effects of a drug can be derived from its ability to change the cellular transcriptome. More than 700 genes were significantly altered in a pan BET inhibitor JQ1-treated hepatocellular carcinoma HepG2 cells 25  Human BET-family proteins contain two bromodomains, each formed by four α helices linked by hydrophobic ZA and BC loops, which surround a conserved acetyl-lysine (KAc)-binding pocket. 6,26,27 Although both BD1 and BD2 of the BET proteins share highly conserved sequences (about 95%) and prefer to engage to a motif with two acetylated lysines bridged by two amino acids (KAc-XX-KAc), dif-  10 We investigated the drug effects in fibroblasts in vitro, as these cells are considered to be among the main players in the cellular response of wound-healing and fibrosis evoked by irradiation. 16,29 Fibroblasts participate in specific steps such as migration to the wound and production of ECM, which can be readily investigated in cultivated fibroblasts from skin or tissues. We focused our analysis on DGKA and ECM proteins as DGKA can control multiple critical steps including the immune response, lipid signaling, cell migration and cell proliferation. 30 DGKA was also reported to control profibrotic gene expression after irradiation through an EGR1-DGKA-COL1A1/3A1 axis in human fibroblasts. 5,6 To further translate our promising data from fibroblasts to clinical applications in fibrosis prevention and treatment, further investigations in preclinical models and clinical trials are, however, required.
Strikingly, we observed considerable variation in the radiation response among fibroblasts with low DNA methylation. We suggest that this observation might reflect the high plasticity and heterogeneity, which fibroblasts require for their multiple physiologic functions in healthy and damaged tissues. Fibroblasts can be trans-activated by various stimuli to increase ECM secretion and remodeling, to foster secretion of signaling factors to surrounding cells, to generate mechanical force, and regulate tissue metabolism and metabolite secretion. 31 During the pathogenesis of fibrosis, this plasticity might be harmful and could even drive the disease. Thus, variation was strongest in the fibroblasts with hypo-methylation at the DGKA enhancer, which was reported to be associated with fibrosis risk. 5 In addition to measuring ECM production, we determined the wound healing potential of our cells by measuring cell migration as an outcome of fibroblast activation response. 32 Our results showed the migration activity was not significantly reduced when treated with BD-selective BET inhibitors at concentrations where the transcription of ECM proteins is already decreased. In pancreatic ductal adenocarcinoma cells; however, the pan-BET inhibitor I-BET762 suppressed migration and invasion. 33 This difference might be caused by the different compounds and cell types tested, but it might also suggest that tumor cells are more sensitive to BET inhibition than normal fibroblasts.
However, cell migration is an important early step which is required in wound healing and tissue regeneration after irradiation. Inhibiting this step would strongly increase side effects after radiotherapy.
Overall, our studies indicate that selective inhibition of the second bromodomain (BD2) of the BET proteins could offer a novel pharmacological approach to prevent the profibrotic response after irradiation. Future and more detailed investigations in different types of fibroblasts are needed to elucidate in more detail how iBET-BD2 affects the radiation-induced profibrotic response. Moreover, extensive preclinical studies are required to verify the efficacy and toxicity of these drugs and whether they can be used to prevent fibrosis development after radiotherapy. AUTHOR CONTRIBUTIONS