Inhibiting BRAF Oncogene–Mediated Radioresistance Effectively Radiosensitizes BRAF^V600E-Mutant Thyroid Cancer Cells by Constraining DNA Double-Strand Break Repair

The RAS–RAF–MEK–ERK signaling cascade is the predominant cell growth–promoting signaling pathway deregulated in most human cancers, and thus a major hub for developing molecularly targeted therapies (1–3). The most frequently mutated component of this pathway is RAS, but directly targeting mutant RAS has proved challenging. Therefore, targeting downstream RAF, MEK, and ERK is being actively pursued in cancer therapeutics (4, 5). RAF kinase family consists of ARAF, BRAF, and CRAF members with distinct mutation patterns in different cancer types (6). BRAF mutations are found in about 8% of all human cancers and are commonly present in melanoma (∼50%), thyroid cancer (∼45%), colorectal cancer (∼10%), non–small cell lung carcinoma (∼5%), and hairy cell leukemia (∼100%; refs. 6–9). The most frequent BRAF mutation is a 1799T>A substitution resulting in constitutively active BRAF^V600E mutant, which leads to stimuli-independent activation of downstream MEK–ERK signaling (6, 9). Thus, specifically targeting the BRAF^V600E mutant is a tumor-selective approach and a major goal of cancer therapy. To that end, several BRAF^V600E-specific small-molecule inhibitors have been developed for the treatment of BRAF^V600E-driving cancers and some have been approved for use in BRAF-mutant melanoma, thyroid cancer, and non–small cell lung cancer (6, 10). However, it has been well documented that BRAF^V600E inhibitors are rarely curative, and the response of patients with cancer to BRAF^V600E inhibitors alone is transient partially due to the paradoxical activation of RAS–ERK signaling through multiple mechanisms (8). The resistance of various cancers to BRAF^V600E inhibitors has led to the development of vertical targeting RAS–ERK signaling by simultaneous inhibition of BRAF^V600E and MEK kinases. Indeed, combinatory BRAF^V600E and MEK inhibitors have shown improved clinical efficacy compared with BRAF^V600E inhibitor as monotherapy and has become the standard therapy for BRAF^V600E metastatic melanoma (11, 12).

Thyroid cancer is the most common cancer of the endocrine system, with an annual incidence of more than 60,000 cases in the United States (13). Thyroid cancers from follicular cells are classified into well differentiated, poorly differentiated, and anaplastic carcinoma. Although well-differentiated thyroid cancers, such as papillary thyroid cancer (PTC), have a good prognosis, the mortality rates of poorly differentiated thyroid cancer (PDTC) and anaplastic thyroid cancer (ATC) are 38%–57% and approximately 100%, respectively (14). Thus, there is an urgent and unmet need to develop novel therapeutic strategies for PDTC and ATC. Recent genomic landscape analyses of thyroid cancers have demonstrated BRAF mutations occur in approximately 60% of PTC, approximately 33% of PDTC, and approximately 45% of ATC, and preclinical studies have established that BRAF or RAS mutations are driver mutations in the development of these thyroid cancers (15–18). Of note, BRAF mutations are mutually exclusive with RAS mutations. Interestingly, combined analyses of gene variance, mRNA expression, protein expression, and phosphorylation of signaling molecules have unraveled the existence of two distinct PTC groups: BRAF^V600E-like and RAS-like. BRAF^V600E-like PTCs have overactivation of MEK–ERK signaling companied with high expression of DUSP genes, while RAS-like PTCs have simultaneous activation of PI3K–AKT–mTOR and MEK–ERK pathways (16).

Thus, BRAF^V600E inhibitors are potential drugs for the treatment of patients with thyroid cancer with BRAF^V600E mutation. As proof of principle, it has been shown that BRAF^V600E inhibitor vemurafenib (FDA approved for the treatment of BRAF-mutant melanoma) improves the clinical outcome of patients with thyroid cancer with BRAF^V600E mutation (3, 19–22). The resistance of thyroid cancer cells to vemurafenib has been ascribed to the rapid reactivation of ERK signaling resulting from vemurafenib-induced expression of ERBB3 and its ligand neuregulin-1 (23). Consequently, it was found that EGFR kinase inhibitor lapatinib effectively prevents MAPK reactivation and sensitizes BRAF^V600E thyroid cancer cells to vemurafenib (24). In addition, it was recently reported that combinatorial administration of BRAF^V600E inhibitor dabrafenib and MEK inhibitor trametinib has robust clinical activity in patients with BRAF^V600E-mutated ATC (3). Thus, rationally designed combinations with targeted drugs, chemotherapy, and radiotherapy may improve the outcome of thyroid cancers with distinct signaling pathways and subvert resistance to drug monotherapy.

DNA damage repair plays an important role in cancer drug resistance (25, 26). RAS mutations have been previously shown to promote preclinical radioresistance and targeted inhibition of MEK can effectively radiosensitize RAS-mutant tumors (27–31). In our recent preclinical study, we found that trametinib attenuates DNA repair pathways in KRAS-mutant pancreatic cancer cells, suggesting that these DNA repair and cell-cycle proteins downstream of MEK–ERK are mediating radioresistance in these tumor cells (28). With regard to BRAF inhibition as a radiosensitizing strategy, there are emerging clinical data suggesting that radiosensitization with BRAF inhibitors can be associated with profound increases in treatment efficacy (32).

Locoregional and systemic tumor control in thyroid cancer, especially PDTC and ATC, are very poor due, in large part, to radiotherapy and chemotherapy resistance. The role of BRAF mutations in radiotherapy resistance is largely unknown. Moreover, whether BRAF^V600E inhibitors (BRAFi) sensitize thyroid cancers with a BRAF^V600E mutation to radiotherapy remains to be determined. In this study, we investigated the role of oncogenic BRAF in radioresistance, and tested radiosensitization of thyroid cancer cells to BRAFi in cell culture and mouse xenograft models. We found that BRAF^V600E thyroid cancer cells were resistant to radiation, and BRAF^V600E mutation resulted in enhanced NHEJ activity leading to radioresistance. In addition, BRAFi selectively radiosensitized BRAF^V600E thyroid cancer cells likely through abrogation of heightened DNA double-strand break repair. Our findings suggest that pharmacologic BRAF^V600E inhibition in combination of radiotherapy may improve the therapeutic outcome of BRAF^V600E-mutant thyroid cancers.

The detailed information of the thyroid cancer cell lines TPC-1, U-Hth-74, SW1736, BCPAP, and 8505C is listed in Supplementary Table S1. We also used A375 and HS294T melanoma cells. Cells were maintained at 37°C in 5% CO₂ in DMEM (TPC-1, Hth-74, A375, and HS294T) or RPMI1640 (SW1736, BCPAP, and 8505C) supplemented with 10% FBS (Sigma), 1% penicillin/streptomycin, and nonessential amino acids (Thermo Fisher Scientific). Cells were cultured for no more than 3 months continuously. Vemurafenib (SelleckChem) was dissolved in DMSO and added to appropriate media with a final concentration of no more than 0.1% DMSO. Total ERK-1/2, phospho-ERK-1/2 (Thr202/Tyr204), total MEK-1/2, phospho-MEK-1/2 (Ser217/221), total BRAF, phospho-BRAF (Ser445), phospho-H2AX, XLF/NHEJ-1, and GAPDH primary antibodies were purchased from Cell Signaling Technology. We purchased BRAF^V600E-specific antibody (26039) from NewEast Biosciences. BRAF^V600E-specific antibody (26039) was from NewEast Biosciences. Anti-rabbit and anti-mouse immunofluorescent secondary antibodies were purchased from LI-COR Biosciences. SMARTpool siGENOME siRNA targeting XLF (NHEJ1) was purchased from Dharmacon Inc. and 25 nmol/L siRNA was transfected into cells by TransIT-X2 Transfection Reagent according to the manufacturer's protocol (Mirus Bio). XLF (NHEJ1) cDNA plasmid was obtained from Genscript (clone OHu23739).

Immunoblotting was performed as described previously (33). Briefly, cell lysates were prepared using RIPA buffer (Thermo Fisher Scientific) supplemented with 1× protease (cOmplete, Roche) and phosphatase inhibitors (PhosSTOP, Roche) followed by protein quantification by the DC Protein Assay Kit (Bio-Rad). Equal amounts of protein were loaded and resolved by SDS/PAGE and transferred onto nitrocellulose membranes. Membranes were incubated in 5% BSA in TBS-Tween blocking buffer for 1 hour at room temperature. Primary antibodies were allowed to bind overnight at 4°C, and used at a dilution of 1:200–1,000. After washing in TBS-Tween, the membranes were incubated with immunofluorescent secondary antibodies at a 1:5,000 dilution for 1 hour at room temperature. Membranes were washed with TBS-Tween and allowed to air dry prior to imaging via LI-COR Odyssey CLx Imaging System.

Cells were trypsinized to generate single-cell suspensions and seeded onto 60-mm or 100-mm tissue culture plates in triplicate essentially as described previously (33). Cells were incubated with DMSO or vemurafenib for 3 hours then irradiated with various doses (0–8 Gy). Twenty-four hours after radiation, medium was replaced with fresh medium without DMSO or vemurafenib. Seven to 10 days after seeding, colonies were fixed with Methanol/Acetic Acid, and stained with 0.5% crystal violet. The numbers of colonies or colony-forming units (CFU) containing at least 50 cells were counted using a dissecting microscope (Leica Microsystems, Inc.) and surviving fractions calculated. Experiments were repeated for multiple, independent times.

AlamarBlue proliferation assay was performed according to manufacturer's instructions (Bio-Rad Antibodies). Briefly, cells were seeded in 96-well plates in six replicates at a density of 1,000–2,000 cells per well in 100-μL medium (day 0). At various days after plating, AlamarBlue reagent was added and incubated at 37°C for 4 hours, and absorbance was measured at 570 and 600 nm.

Irradiation was performed essentially as described previously with 160 kV, 25 mA at a dose rate of approximately 113 cGy/minute using a RS-2000 biological irradiator (RadSource; ref. 31).

TPC-1 cells were transfected with empty pBabe-puro plasmid (#1764, Addgene) or pBabe-puro-BRAF-V600E plasmid (#15269, Addgene) using TransIT-2020 Transfection Reagent as per manufacturer's protocol (Mirus Bio). Cells were seeded in 6-well plates at a density of 2.5 × 10⁵ cells/mL and cultured overnight. To prepare the TransIT-2020:DNA complexes, 250 μL of OptiMEM I Reduced-Serum Medium, 2.5 μL of plasmid DNA (1 μg/μL), and 7.5 μL of TransIT-2020 were mixed in a sterile tube by pipetting. The complex mixtures were allowed to incubate at room temperature for 30 minutes and then added to the cells. After 48 hours of incubation, cells were harvested and cultured overnight. Cells were treated with 1.0 μg/mL of puromycin to select for stable clones. Single stable clone colonies were picked using sterile cloning cylinders and seeded separately in 12-well plates. Each clone was cultured and expanded once confluent. Stable transfection was verified by immunoblotting using a BRAF^V600E-specific antibody.

Cells were treated with DMSO or vemurafenib for different time points. Neutral comet assay was performed as described previously (28).

For γH2AX, RAD51, and 53BP1, immunofluorescence for nuclear foci was performed as described previously (28).

Using TransIT-2020 Transfection Reagent (Mirus Bio), 293T cells were transiently transfected either empty pBabe-puro plasmid or pBabe-puro-BRAF-V600E plasmid. Cells were cotransfected with pDR-GFP plasmid previously characterized (34) or NHEJ-GFP-PEM1 plasmid (generously provided by V. Gorbunova) to measure homologous recombination (HR) and nonhomologous end-joining (NHEJ) activity, respectively. All cells were also transfected with pLVX-mCherry-C1 as a transfection control. Cells were treated with adenovirus expressing the I-SceI restriction enzyme that induces double-strand DNA breaks in the recognition sequence within the reporter construct. Cells were also pretreated with DMSO or vemurafenib using different variations in scheduling as described. Then, reporter assays were performed as described previously (28). Repair efficiency was expressed as the ratio of percent GFP⁺ cells over percent mCherry⁺ cells.

Animal studies were conducted in accordance with an approved protocol adhering to the Institutional Animal Care and Use Committee policies and procedures at The Ohio State University (Columbus, OH). Six- to 8-week-old male athymic nude mice (Taconic Farms Inc.) were caged in groups of five or less, and fed a diet of animal chow and water ad libitum. 8505C and BCPAP cells were injected subcutaneously into the flanks of each mouse at 2 × 10⁶ and 2.5 × 10⁶ cells per injection, respectively. Treatment regimens were started once tumors reached approximately 150 mm³ in size, 2–4 weeks postinjection. Vemurafenib powder (Active Biochem) was suspended in 0.9% sodium chloride containing 5% dextrose. Vemurafenib was administered orally using a sterile 18G gavage needle at 50 mg/kg twice daily for 5 consecutive days. Using a custom shielding apparatus to block nontargeted areas, 4 Gy of radiation was administered directly to tumors once daily for 5 consecutive days. For combination treatment, mice were treated with radiation 2–3 hours after the daily dose of vemurafenib. To obtain a tumor growth curve, perpendicular diameter measurements of each tumor were measured every 2–5 days from the first day of injection with digital calipers, and volumes were calculated using the formula (L × W × W)/2.

For in vitro experiments, data are presented as the mean ± SEM for clonogenic survival and foci experiments. Statistical comparisons were made between the control and experimental conditions using the two-sided two-group t tests with significance assessed at P < 0.05.

The BRAF mutation status was downloaded from http://www.cbioportal.org/ under the mutation data of well-differentiated thyroid cancer [The Cancer Genome Atlas (TCGA), THCA Provisional] named “Tumor samples with mRNA data (RNASeq V2; n = 509). The RSI was calculated on the basis of the following formula using Thyroid cancer TCGA mRNA (RNASeq V2) gene expression data:

Note, the radiosensitivity index (RSI) value predicts the survival fraction at 2 Gy; thus, it is inversely correlated with a tumor's radiosensitivity (i.e., lower RSI value with higher radiosensitivity to external radiation). The correlation between BRAF^V600E mutation and RSI was tested with two-sample t test. Among TCGA THCA tumor samples, 235 had BRAF^V600E mutation and 261 had wild-type (WT) BRAF, so data from 496 patients was used to test the correlation between BRAF mutation and RSI.

Data for BRAF mutations and XLF expression was downloaded from UCSC Xena data portal (https://xena.ucsc.edu/ accessed January 18, 2018) for primary tumor samples from TCGA THCA dataset. XLF expression was reported as log₂ (normcount+1). Patients with WT and BRAF^V600E mutations were included (patients with other BRAF mutations were eliminated). The correlation between XLF expression and BRAF mutation was tested with a linear model. Data were available for 288 patients with BRAF^V600E mutation and 203 WT BRAF tumors for which XLF expression was available.

To assess whether BRAF^V600E mutation is correlated with radioresistance of thyroid cancer cells, the cells of BRAF WT (BRAF^WT: U-Hth-74, TPC-1) or V600E mutation (BRAF^V600E: 8505C, BCPAP, SW-1736) thyroid cancer cell lines were treated with increasing doses of ionizing radiation (IR) followed by colony formation (clonogenic) assay. Compared with BRAF^WT cell lines, there was a relatively higher surviving fraction of thyroid cancer cells with BRAF^V600E in response to IR (Fig. 1A). To confirm this observation, stable TPC-1 cells expressing BRAF^V600E (V600E) or empty vector control (EMP) were established (Western blot inset, Fig. 1B), and subjected to radiation clonogenic assay. We found that expression of BRAF^V600E in BRAF WT TPC-1–stable cells led to marked radioresistance (Fig. 1B), without significant changes in proliferation (Supplementary Figs. S1 and S2). Finally, we evaluated whether BRAF^V600E thyroid tumors in the TCGA database are predicted to be radioresistant using a validated RNA-based molecular signature predicting relative radiosensitivity across tumor cells, the RSI (35, 36), where higher RSI is associated with radioresistance. In support of our preclinical data, BRAF^V600E thyroid cancers are generally predicted to be more radioresistant than BRAF WT thyroid cancer (Fig. 1C and D). The mean calculated RSI for tumors with BRAF^V600E mutation (n = 235) is 0.7216 ± 0.1324, while the mean calculated RSI for tumors with WT BRAF (n = 261) is 0.6945 ± 0.1264 with P = 0.0203. Furthermore, we examined radiation sensitivity data from a collection of over 500 cell lines from Yard and colleagues (37), and pooled radiation sensitivity data from the top three BRAF^V600E-expressing tumor types that contained more than one BRAF-mutant cell line (thyroid, melanoma, and large bowel/colorectal). We found that BRAF^V600E-mutant cell lines were associated with radioresistance compared with cells lacking BRAF^V600E (Supplementary Fig. S3). Taken together, these results support that BRAF^V600E mutation contributes to radioresistance of thyroid cancer cells (and potentially other tumor types).

Vemurafenib is a potent BRAF^V600E-mutant kinase–specific inhibitor (BRAFi) and inhibits ERK signaling in BRAF^V600E-mutant cells with minimal effect in BRAF^WT cells (10). Consistently, vemurafenib reduced pMEK and pERK1/2 in 8505C and BCPAP cells with BRAF^V600E mutation in a concentration-dependent manner after 3 hours of treatment (Fig. 2A). In contrast, treatment of BRAF^WT TPC-1 and U-Hth-74 cells with vemurafenib did not alter the phosphorylation of MEK (pMEK) and ERK1/2 (pERK1/2) except at perhaps higher doses of BRAFi (e.g., mild decrease at 500 nmol/L), consistent with sustained MEK and ERK activation. We postulated that vemurafinib would sensitize BRAF^V600E thyroid cancer cells to radiotherapy by reversing BRAF^V600E-induced radioresistance. To test this hypothesis, BRAF^WT U-Hth-74 and TPC-1 cells, and BRAF^V600E 8505C and BCPAP cells were pretreated with 100 nmol/L BRAFi for 3 hours and then exposed to different doses of IR, followed by radiation clonogenic assays. We found that BRAFi treatment significantly sensitized BRAF^V600E 8505C and BCPAP cells, but not BRAF^WT U-Hth-74 and TPC-1 cells to IR (Fig. 2B; Supplementary Fig. S4). Differences in survival were unlikely related to toxicity of vemurafenib, as plating efficiency/colony formation in the absence of radiation was virtually unaffected by vemurafenib (Supplementary Fig. S5). We extended our analysis to other BRAF^V600E-mutant cell lines, A375 and HS294T melanoma cells, and likewise showed that BRAFi caused marked radiosensitization (Supplementary Fig. S6). Our results showed that inhibition of BRAF^V600E-mutant kinase with vemurafenib reversed the radioresistance of thyroid cancer (and melanoma) cells harboring a BRAF^V600E mutation.

Ionizing radiation (IR) induces DNA double-strand breaks (DSBs), which result in rapid activation of the ATM-CHK2 checkpoint to arrest the cell cycle and promote repair machineries to restore the damaged DNA. If left unrepaired, IR-induced DSBs result in cell death (25, 26). Phosphorylation of histone H2AX (γH2AX) is a marker of DNA damage and its dynamics reflects the processes of DNA damage production and repair. To assess the role for BRAF^V600E mutation in DNA damage response (DDR) in thyroid cancer cells, 8505C and BCPAP cells were pretreated with 100 nmol/L BRAFi for 3 hours and then exposed to 4 Gy IR. At different time points after IR, γH2AX were assessed by immunoblotting. In both 8505C and BCPAP cells without BRAFi treatment, there was strong γH2AX induction as expected at 15 minutes after IR, and γH2AX expression recovered to basal levels between 6 and 24 hours after IR, indicative of the repair of IR induced DNA damage. However, under the presence of BRAFi, there was persistence of γH2AX expression even 24 hours after IR, indicative of persistent presence of DNA damage (Fig. 3A). To confirm this observation, we performed immunofluorescence analysis of γH2AX nuclear foci of these cells after IR in the presence or absence of BRAFi. As expected, γH2AX foci were formed in the nuclei of 8505C and BCPAP cells soon after radiation (15 minutes). γH2AX signals gradually decreased with time, indicative of recovery from DNA damage, but there were significantly higher levels of γH2AX foci in cells treated with BRAFi than those without BRAFi at 6 and 24 after IR (Fig. 3B). Conversely, changes in γH2AX were absent in BRAF^WT cells in similar immunoblotting and immunocytochemistry assays (Supplementary Fig. S7).

In addition, comet assay is a sensitive technique for detecting DNA damage at the level of individual cells (38). To test whether BRAFi alters the global DNA repair process in irradiated thyroid cancer cells, neutral comet assays were conducted in 8505C cells after 10 Gy IR in the presence or absence of BRAFi. At each time point after IR, there was persistently higher tail moment in the cells treated with BRAFi than those without BRAFi, suggestive of attenuated DNA repair (Fig. 3C). Taken together, these results demonstrate that inhibition of BRAF^V600E delays global DNA repair in BRAF^V600E-mutant thyroid cancer cells.

DNA DSBs are repaired by multiple DNA repair pathways, including nonhomologous end joining (NHEJ) and homologous recombination (HR). Enhanced activity of NHEJ and HR repair pathways play an important role in the resistance of cancer cells to radiotherapy and DNA damage–based chemotherapy (25, 26). Moreover, deregulation of cell growth signaling pathways have been shown to maintain cancer cell survival by promoting DNA damage repair at the cost of increased mutation risk (39–41). DSBs induce a rapid colocalization of p53-binding protein 1 (53BP1) with γH2AX and other DNA repair proteins including RAD51 (a component of HR repair), which can be visualized as discrete nuclear foci (39–41). We determined the effects of BRAFi on radiation-induced 53BP1 and RAD51 foci formation in 8505C and BCPAP cells. Numerous 53BP1 foci were formed in the nuclei of 8505C and BCPAP cells after IR, and these foci were resolved with time as expected; however, in the presence of BRAFi, the decrease of 53BP1 foci was significantly impaired during the recovery phase after IR, that is, at 6 and 24 hours after treatment (Fig. 4A). Nuclear RAD51 foci increased after IR up to 6–24 hours after IR as expected. However, in the BRAFi-treated cells, there was significantly increased number of RAD51 foci particularly at 6 and 24 hours (Fig. 4B). To more directly determine whether BRAFi induces radiosensitivity of BRAF^V600E cells by altering HR and/or NHEJ repair, fluorescent reporter constructs that allow sensitive and quantitative measurement of HR (DR-GFP) or NHEJ repair (Pem1-NHEJ) were transfected in 293T cells. These constructs contain an engineered GFP gene with recognition sites for I-SceI endonuclease for induction of DSBs, which does not express GFP in the absence of repair of these I-SceI–induced DSBs. Proper repair of DSBs in the recognition sites results in GFP expression, and the number of GFP-positive cells can be counted by flow cytometry (42). We cotransfected 293T cells with either Pem1-GFP or DR-GFP reporters, with (empty vector) or without BRAF^V600E, and tested these cells with or without BRAFi to determine the effects of BRAF^V600E inhibition on HR and NHEJ repair capability. Expression of BRAF^V600E and subsequent MEK-ERK activation is shown in Supplementary Fig. S8. We found that expression of BRAF^V600E did not significantly alter HR repair (Fig. 4C, middle), and BRAFi only slightly reduced HR repair (but without statistical significance). In the NHEJ repair assay, however, we found that expression of BRAF^V600E significantly increased NHEJ-mediated repair, which was significantly attenuated by BRAFi (Fig. 4C, right), suggesting that BRAF^V600E mutation imparts accelerated DNA repair through an increase in NHEJ repair activity.

To explore the potential mechanisms of the upregulation of NHEJ by BRAF^V600E mutation, we analyzed the thyroid cancer (THCA) TCGA database for the gene expression status of the components of NHEJ and the correlation with BRAF^V600E mutation. We found higher levels of XLF RNA expression in the presence of a BRAF^V600E mutation (7.28 vs. 7.06 log normalized counts, fold change = 1.17; P = 2.49e-14; Fig. 5A), as well as other NHEJ-associated genes (Supplementary Table S2). To more directly determine whether BRAF^V600E affects XLF expression, we compared XLF protein expression by immunoblotting between BRAF^WT (TPC-1 and Hth-74) and several BRAF^V600E (8505C, BCPAP, SW1736) thyroid cancer cell lines, and observed an increase of XLF in BRAF^V600E thyroid cancer cells at both the protein and mRNA levels (Fig. 5B). In support of our findings, treatment of BRAF^V600E thyroid cancer cells with BRAFi led to reduction of XLF mRNA and protein (Fig. 5C). As shown previously, expression of BRAF^V600E in BRAF^WT TPC-1 cells led to radioresistance in clonogenic assay (Fig. 1C). Ectopic expression of BRAF^V600E in these cells resulted in upregulation of XLF (Fig. 5D). Furthermore, the radioresistance induced by BRAF^V600E in TPC-1 cells could be attenuated by BRAFi (Fig. 5D). Consistent with pharmacologic inhibition of BRAF^V600E, genetic silencing of XLF by RNA interference (siRNA) effectively radiosensitized TPC-1-BRAF^V600E cells with minimal effect on TPC-1 parental cells (Fig. 5E). Importantly, knockdown of XLF by siRNA in BRAF^V600E 8505C cells also resulted in radiosensitization (Fig. 5F). Furthermore, we transfected XLF cDNA into BRAF WT TPC-1 cells and assessed radiosensitivity with a radiation clonogenic assay. We found that XLF expression was sufficient to confer increased radioresistance (Supplementary Fig. S9). These results suggest that upregulation of XLF may at least in part contribute to the enhancement of NHEJ-mediated repair of DSBs and concomitant radioresistance of BRAF^V600E thyroid cancer cells.

To translate our findings, we tested whether BRAF inhibition could radiosensitize BRAF^V600E thyroid cancer cells. When tumors of 8505C and BCPAP cell lines reached about 100–200 mm³, the mice were randomized to be treated with vehicle or BRAFi (vemurafenib delivered by oral gavage, 50 mg/kg × 5 consecutive days), and/or radiation (4 Gy × 5 consecutive days). In the group of mice treated with concurrent BRAFi and radiation, BRAFi was dosed 2–3 hours before radiation. BRAFi alone led to tumor growth delay, but resulted in rapid regrowth of tumors after discontinuation of BRAFi. Radiation was effective in halting tumor growth for a period of time, but tumors almost universally recurred. However, BRAFi in combination with radiation led to marked and sustained regression of the tumors in multiple BRAF^V600E-mutant xenograft models (log-rank P < 0.0001; Fig. 6A and B), and longer tumor-doubling rates (log-rank P = 0.0057; Fig. 6C and D). In terms of tolerance, treatment with BRAF inhibitor did not result in significant weight loss (Supplementary Fig. S10). While mice treated with radiation (with or without BRAFi) experienced up to 7%–8% weight loss 1 week within the start of radiation, they recovered to their baseline weight by approximately 2 weeks after therapy was complete, and no significant clinical toxicity was noted.

In this study, we demonstrate that for the first time that BRAF^V600E directly promotes radiation resistance through heightened nuclear DNA DSB repair, and that a potent BRAF inhibitor suppresses the resistance of BRAF^V600E thyroid cancer cells to radiotherapy. We further showed that BRAF^V600E thyroid cancer cells displayed an increase of NHEJ activity with upregulation of XLF at the protein and RNA level, a key component of NHEJ. Importantly, BRAF inhibition in combination with radiotherapy led to a marked and sustained tumor regression of BRAF^V600E thyroid tumor xenografts. Our studies provide the basis for clinical testing of combinatorial radiotherapy and BRAF inhibitor treatment modalities for thyroid cancers, particularly BRAF^V600E-mutated ATC.

Another study has demonstrated the efficacy of combining BRAF inhibitors in combination with radiation for BRAF-mutant tumors. Dasgupta and colleagues showed in high-grade glioma preclinical models that BRAFi and radiotherapy showed greater antitumor effects than either treatment alone in BRAF^V600E (but not BRAF wild-type) lines (43). They noted that the combination treatment increased apoptotic cell death, decreased Ki-67 and phospho-MAPK signaling, and increased γH2AX compared with control tumor cells. Our study joins this one in demonstrating the efficacy of combined radiotherapy and BRAFi for BRAF^V600E tumors.

Among the different subtypes of thyroid cancer, ATC has the worst prognosis with an overall survival at 2 years of 10% and mortality of approximately 95%–100%, making it one of the most lethal solid tumors (44). Recent genomic profiling of ATC has revealed high-frequency mutations of the RAS-ERK signaling pathway, which is well-characterized as important for tumor cell proliferation, aggressiveness, and response to therapy (15). Oncogenic mutations in BRAF are the most common among the RAS-ERK pathway genes, occurring in approximately 45% of ATC with the vast majority representing BRAF^V600E-activating mutations (>95%; refs. 14, 15, 17, 18). BRAF mutations are also commonly found in differentiated thyroid cancers, particularly in 60% of PTC, and have been linked to radioactive iodine resistance (16, 45–47). In addition, RAS mutations occur in an additional 20%–25% of patients with ATC (NRAS 15%–20%, HRAS 5%, KRAS 5%), and are mutually exclusive with BRAF mutations (14, 15, 18). Thus, BRAF represents an attractive target for BRAF-mutant thyroid cancers. However, clinical trials have shown that BRAF^V600E inhibitors as monotherapy, including BRAFi, are not curative for patients with cancer with BRAF^V600E, and resistance emerges due to paradoxical activation of RAS-ERK signaling (19–22). Even though vertical targeting of RAS-ERK bypass resistance mechanisms through dual targeting of BRAF and MEK kinases improves the survival of patients with cancer with BRAF^V600E mutation, the tumors still relapse at least partially due to paradoxical RAS-ERK activation, similar to tumors treated with BRAF inhibitors alone (6). Our preclinical data suggest that BRAF^V600E promotes intrinsic radioresistance in thyroid cancer cells and that subversion of oncogene-mediated radioresistance leads to the marked efficacy of the combination of BRAFi and IR in BRAF^V600E xenograft models, with profound delays in tumor growth without significant toxicity. Regarding that radioresistance is one of the major mechanisms of the high rate of disease failure of ATC, our findings suggest that targeting BRAF^V600E in parallel with radiation is a novel strategy for the management of patients with ATC.

RAS mutations have been previously shown to promote preclinical radioresistance in various tumor types, and targeted inhibition of downstream MEK-1/2 can effectively radiosensitize RAS-mutant tumors as shown by our group and others (27–31). Furthermore, RAS mutations have recently been associated with clinical radioresistance in multiple tumor types (48, 49). In this study, we found that BRAF^V600E promotes intrinsic radioresistance in thyroid cancer cells through accelerated NHEJ repair activity, which was prevented by BRAFi in both BRAF^V600E mutation thyroid cancer cell lines and an isogenic thyroid cancer cell line stably expressing BRAF^V600E. Together, it seems that upregulation of DNA repair capacity by oncogenic proteins such as BRAF and RAS mutations may be a common mechanism by which deregulation of cell growth and survival pathways promotes not only tumorigenesis but therapeutic resistance.

Ionizing radiation causes DSBs, which result in mitotic catastrophe and cell death if left unrepaired (25, 26). DSBs are repaired by NHEJ as well as HR. HR repair is restricted mostly to S and G₂ phases of the cell cycle, and is a higher fidelity process. NHEJ repair is more error-prone and operates throughout the cell cycle. NHEJ is activated by Ku70/Ku80 sensor proteins, which subsequently activate DNA-PK catalytic subunit to recruit effector proteins to DSB sites, including XLF (NHEJ1; refs. 25, 26). XLF, in turn, interacts with XRCC4 and DNA ligase IV and may be involved in the end-bridging or ligation steps of NHEJ. It has been recently established that BRAF and RAS mutations contribute to tumorigenesis and tumor progression of thyroid cancers (14). However, the role of BRAF and RAS mutations in radiotherapy and other genotoxic therapy response is largely unknown, particularly for BRAF mutations. We revealed that BRAF^V600E mutation led to radiation resistance through up-regulation of NHEJ, which may, in part, be mediated by XLF upregulation, and XLF expression was decreased by a BRAF^V600E inhibitor. Thus, the upregulation of XLF-mediated NHEJ repair may be a contributing factor to the radioresistance observed in BRAF^V600E thyroid cancer cells, and inhibition of XLF or other NHEJ activity may represent alternative therapeutic strategies. Nevertheless, in the future, it will be important to more thoroughly elucidate the molecular mechanisms of the upregulation of XLF by BRAF^V600E mutation in thyroid cancer as well as whether this is an important mediator of radioresistance in other RAS- and BRAF-mutated cancers (e.g., melanoma, colorectal cancer, non–small cell lung cancer, and pancreatic cancer).

The discovery of oncogene addiction of a subset of cancers to BRAF mutations has triggered the development of a wave of RAF kinase inhibitors for cancer therapeutics. More recently, the rapid appearance of RAF kinase inhibitor resistance has led to the combination of RAF inhibitors with other targeted drugs, and development of third-generation RAF kinase inhibitors (6). Multiple mechanisms contribute to the resistance to BRAF^V600E inhibitors, including paradoxical activation of RAS-ERK signaling as mentioned above (50–52). Consequently, it has been shown that vertically targeting RAS-ERK signaling by simultaneous inhibition of BRAF^V600E and MEK kinases improves clinical efficacy of both thyroid and melanoma with BRAF^V600E in comparison with BRAF^V600E inhibitors as monotherapy (3, 11, 12). In addition, activation of PI3K–AKT–mTOR signaling and modulation of the cell apoptosis cascade play important roles in resistance to BRAF^V600E inhibitors (50–52). One of the pivotal roles of PI3K–AKT–mTOR signaling is the control of protein translation via regulating eIF4F complex formation. Intriguingly, it was observed that the persistent formation of eIF4F complex contributes to BRAF^V600E inhibitors' resistance (53). It is possible that the combination of BRAF^V600E inhibitors with PI3K–AKT–mTOR signaling inhibitors or apoptosis agonists such as IAP antagonists may reverse the resistance of BRAF^V600E-driving cancers to BRAF^V600E inhibitors, and thereby improve overall clinical outcomes.

In summary, we demonstrate for the first time that BRAF^V600E promotes the radioresistance of thyroid cancer cells by enhancing NHEJ-mediated DSB repair activity, which may be due, in part, to XLF upregulation. More importantly, BRAF^V600E radioresistance is abrogated by BRAF^V600E inhibitor both in cell culture and mouse xenografts. Our findings suggest that upregulation of DNA damage repair is one of the mechanisms by which BRAF^V600E exerts therapeutic resistance. Taken together, these results provide strong rationale for clinical testing of radiotherapy and BRAF inhibitors for BRAF^V600E-mutant thyroid cancers.

No potential conflicts of interest were disclosed.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Advancing Translational Sciences or the NIH.

Conception and design: R. Robb, L. Yang, C. Shen, M.D. Ringel, T.M. Williams

Development of methodology: R. Robb, L. Yang, S.M. Jhiang, T.M. Williams

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Robb, L. Yang, A.R. Wolfe, M. Vedaie, S.M. Jhiang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Robb, L. Yang, C. Shen, A.R. Wolfe, A. Webb, X. Zhang, S.M. Jhiang, T.M. Williams

Writing, review, and/or revision of the manuscript: C. Shen, A.R. Wolfe, A. Webb, X. Zhang, M. Saji, S.M. Jhiang, M.D. Ringel, T.M. Williams

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Vedaie, M. Saji, T.M. Williams

Study supervision: T.M. Williams

This work was supported by the following grants: American Cancer Society (RSG-17-221-01-TBG, to T.M. Williams), NIH P50CA168505 (Thyroid SPORE CDP, to T.M. Williams), and National Center for Advancing Translational Sciences (KL2TR001068, to T.M. Williams). Research reported in this article was also supported by The Ohio State University Comprehensive Cancer Center (OSU-CCC) and NIH (P30 CA016058).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.