Distribution and Activity of Lenvatinib in Brain Tumor Models of Human Anaplastic Thyroid Cancer Cells in Severe Combined Immune Deficient Mice

Thyroid cancer is one of the most common solid tumor types, with an increased incidence over the last decade in both the United States and Japan (1, 2). The prognosis of thyroid cancer varies widely among histologic types. In general, differentiated thyroid cancer (DTC), which accounts for approximately 90% of thyroid cancers, consist of papillary and follicular thyroid carcinoma and shows a gradually progressive clinical course, with an overall survival time >10 years of standard surgery and radioactive iodine (RAI) therapy (1). Conversely, anaplastic thyroid cancer (ATC), which accounts for approximately 1%–2% of all thyroid cancers, shows a very aggressive clinical course, with most accompanied by extra-glandular lesions, including distant metastases, at the time of diagnosis (3, 4). ATC, which is considered stage IV by American Joint Committee on Cancer criteria, is refractory to standard treatments for thyroid cancer, such as RAI and chemotherapy (5). Thus, its prognosis is extremely poor, and the survival rate of patients with ATC 6 months after the initial diagnosis is less than 10%.

Brain metastases often develop in many types of cancers, including lung, breast, and renal cancer, as well as approximately 15% of patients with ATC (6). Brain metastases are considered a lethal complication that shortens survival and markedly diminishes patient quality of life (7). Although palliative surgery and radiotherapy, including whole-brain irradiation (WBI) and stereotactic radiotherapy are available, their clinical benefits are limited to a subpopulation of patients with good functional status, such as younger patients with controlled primary tumors (8). More effective treatment methods are therefore needed for patients with brain metastases of many types of cancer, including ATC.

Recently, three orally available targeted drugs, sorafenib, vandetanib, and lenvatinib, were approved for different types of thyroid cancer. Vandetanib was approved for medullary thyroid cancer (9), and sorafenib and lenvatinib were approved for DTC worldwide (10–12). In addition, lenvatinib has been approved for ATC in Japan (13). A recent phase III clinical trial (selenium and vitamin E cancer prevention trial) demonstrated that progression-free survival in patients with iodine-131-refractory differentiation thyroid cancer was longer for those who received lenvatinib than placebo (14). While treatment-related adverse events (AE) occurred for almost of all patients, the patients who experienced serious treatment-emergent AEs were similar in younger and older patients in the lenvatinib arm. The most common treatment-related AEs were hypertension, diarrhea, decreased appetite, and decreased weight (14). These drugs are multiple kinase inhibitors; sorafenib is an inhibitor of VEGFRs, platelet-derived growth factor receptors (PDGFR), RET/PTC, BRAF, and c-Kit (15); vandetanib inhibits VEGFRs, EGFRs, and RET/PTC (16); and lenvatinib inhibits VEGFRs, FGFRs, PDGFRα, RET/PTC, and c-Kit (17). Preclinical studies showed that both sorafenib and lenvatinib suppress angiogenesis and thereby inhibit the growth of thyroid cancer cells, including ATC cells, in xenograft mouse models (18, 19). However, the effects of these compounds against ATC brain metastases remain undetermined. We have encountered a patient with DTC whose brain metastases were successfully controlled by lenvatinib (Fig. 1A). Thus, this study assessed the effects of lenvatinib, compared with sorafenib, against brain metastases of ATC and explored the mechanisms underlying the effects of lenvatinib in preclinical models.

A 70-year-old man with brain metastases of advanced papillary thyroid carcinoma was treated with lenvatinib. During treatment, this patient underwent brain CT examined approximately every 3 months. Plane or contrast-enhanced brain CT axial view at 5-mm slice thickness showed tumors and edema around the tumors in right frontal lobe before treatment with lenvatinib. Treatment with lenvatinib led to a decrease in tumor size and the regression of edema around the tumors for longer than 4 months. The study was conducted according to the Declaration of Helsinki. The patient participated in the Institutional Review Board of Kanazawa University–approved study (No. 2016-222) and provided written informed consent.

Six human ATC cell lines, OCUT-1C, OCUT-1F, OCUT-2, OCUT-4, OCUT-5, and OCUT-6, which had been established from surgical specimens obtained from Japanese patients (20), were maintained in RPMI1640 medium, supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (50 mg/mL) in a humidified CO₂ incubator at 37°C. All cells were passaged for <3 months before renewal from frozen, early-passage stocks. Human dermal microvascular endothelial cells (HMVEC) were maintained in HuMedia-MvG with growth supplements (Kurabo) and used for in vitro assays at passages 2–5. Sorafenib was obtained from Selleck Chemicals and lenvatinib was from Eisai Co., Ltd.

Cell viability was measured using the MTT [3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium] dye reduction method (21). Tumor cells plated at a density of 2 × 10³ cells in 100 μL RPMI1640 plus 10% FBS per well in 96-well flat or square plates, were incubated for 24 hours. Sorafenib or lenvatinib was added to each well, and incubation was continued for another 72 hours. Cell growth was measured using MTT solution (2 mg/mL; Sigma).

Protein aliquots (20 μg) were separated by SDS-PAGE (Bio-Rad) and transferred to polyvinylidene difluoride membranes (Bio-Rad). The membranes were washed three times and incubated with Blocking One Blocking Reagent (Nacalai Tesque, Inc.) for 1 hour at room temperature. After washing, the membranes were incubated overnight at 4°C with primary antibodies against PDGFRα (D13C6), phospho-PDGFRα (Y762), FGFR (D814), phospho-FGFR1 (Y653/654), VEGFR2, phospho-FRS2α (Y436), FRS2α (Y436), cleaved PARP, Cyclin A2 (BF683), β-actin (Cell Signaling Technology), and cyclin A (H432; Santa Cruz Biotechnology). The membranes were washed and incubated for 1 hour at room temperature with species-specific horseradish peroxidase–conjugated secondary antibodies. Immunoreactive bands were visualized using an enhanced chemiluminescent substrate (SuperSignal West Dura Extended Duration Substrate, Pierce Biotechnology). All data are representative of three independent experiments.

Cells (2 × 10⁵) were cultured in 2 mL of RPMI1640 with 10% FBS for 24 hours, washed with PBS, and incubated for 48 hours in RPMI1640 with 10% FBS. The culture media were harvested and centrifuged, and the supernatants were stored at −80°C until analysis. The indicated cytokines in the culture media were measured by Quantikine ELISA Kits (R&D Systems), according to the manufacturer's instructions. All samples were assayed in triplicate. Color intensity at 450 nm was measured with a spectrophotometric plate reader.

Suspensions of 5 × 10⁶ OCUT-1C and OCUT-2 cells each were injected into the backs of 5-week-old male SHO-Prkdc^scidHr^hr mice (SHO mice, Charles River). When tumor volume reached approximately 100 mm³, calculated as mm³ = width² × length/2, the mice were randomized at into groups of 5. Lenvatinib (10 mg/kg/d or 50 mg/kg/d), sorafenib (25 mg/kg/d or 50 mg/kg/d), or saline was administered once daily by oral gavage. Tumor volume and mouse body weights were measured twice weekly. All animal experiments complied with the Guidelines of the Institute for Experimental Animals, Kanazawa University Advanced Science Research Center (approval number: AP-081088; Kanazawa, Japan).

EGFP cDNA was excised from the vector pIRES-EGFP (Invitrogen) and Emerald Luc (Eluc) was excised from the vector Emerald Vector (ELV-101; Toyobo). Both fragments were cloned into the MaRXIVf Puro retroviral vector (22). All cDNAs were sequenced. Vectors were transfected into mammalian cells using Lipofectamine LTX and PLUS Reagent (Invitrogen), and retroviral infection was performed using an Ampho Retrovirus Packaging Kit (Takara Bio Inc.) according to the manufacturers' instructions, with transfected cells selected with puromycin.

Mouse scalps were sterilized with 70% ethanol, and a small hole was bored into each skull, 0.5 mm anterior and 3.0 mm lateral to the bregma, using a dental drill. Suspensions of OCUT-1C/EGFP-Eluc cells (1.5 × 10⁵/1.5 μL) were injected into the right striatum, 5 mm below the surface of the brain, using a 10 μL Hamilton syringe with a 26-G needle. The scalp was closed using autoclips (BD). After 9 days, the mice were randomized at into groups of 4 and treated orally with saline, sorafenib (50 mg/kg/d), or lenvatinib (10 mg/kg/d). Mouse body weights were measured twice weekly. All animal experiments were in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the Ministry of Education, Culture, Sports, Science and Technology, Japan. The protocol was approved by the Committee on the Ethics of Experimental Animals, and the Advanced Science Research Center, Kanazawa University (Kanazawa, Japan; approval no. AP-132618). All surgeries were carried out under sodium pentobarbital anesthesia, and efforts were made to minimize the suffering of the animals.

After inoculation, tumor development was tracked in live mice by repeated noninvasive optical imaging of tumor-specific luciferase activity using the IVIS Lumina XR Imaging System (PerkinElmer), as described previously (23). The intensity of the bioluminescence signal was analyzed using Living Image 4.0 Software (PerkinElmer) by serially quantifying the peak photon flux in a selected region of interest (ROI) within a given tumor. The intensity of the bioluminescence signal was corrected to take into account the total area of the ROI, and the elapsed time during which bioluminescence signals were read by the CCD camera. The corrected value was expressed as photons/s.

Formalin-fixed, paraffin-embedded tissue sections (4-μm thick) were deparaffinized. Proliferating cells were detected by incubating tissue sections with anti-Ki-67 antibody (Clone MIB-1; Dako Corp). Microvessel density was assessed using anti-CD-31 antibody (ab28364; Abcam). Antigens were retrieved by microwaving tissue sections in 10 mmol/L citrate buffer (pH 6.0) and Tris-EDTA buffer (pH 9.0), respectively. After incubation with a secondary antibody and treatment with the Histofine Simple Stain MAX-PO (R) Kit (Nichirei), peroxidase activity was visualized via a DAB reaction. The sections were counterstained with hematoxylin. All sections were stained with hematoxylin and eosin (H&E) for routine histologic examinations.

OCUT-1C cells were inoculated into the brains of 9 mice, with chemiluminescence observed in all 9 after 40 days. Mice were treated with water (control, N = 2), lenvatinib (10 mg/kg, N = 4), or sorafenib (50 mg/kg, N = 3) for 3 days. In a parallel experiment, we inoculated OUCT-1C cells into the subcutaneous spaces of 9 mice and treated these mice with water (control), lenvatinib (10 mg/kg), or sorafenib (50 mg/kg; N = 3/group). Brain and subcutaneous tumors were harvested. Frozen tissue sections (8-μm thick) fixed in cold acetone were stained with anti-CD 31 antibody. After blocking with 5% FBS, the samples were stained with anti-CD-31 mAb (MEC13.3; BD PharMingen) and goat-anti-rat IgG conjugated to Alexa Fluor 594 (Invitrogen). The slides were mounted by Vectashield HardSet (Vector Laboratories). Localized blue and red fluorescence was detected using a fluorescence microscope.

The five areas within a section showing the highest levels of staining intensity were selected for histologic quantification by light or fluorescence microscopy at 200-fold magnification. All results were independently evaluated by two authors (T. Yamada and R. Wang).

Mouse brains were dissected using surgical scissors and tweezers at room temperature, flash-frozen in liquid N₂, and stored at −80°C until use. Brain tissue was embedded into super cryo-embedding medium (Section Lab Co. Ltd.) and cut into serial coronal sections (16 μm) using a cryostat (CM-3050 S; Leica Microsystems) at −25°C. The sections were thaw-mounted on indium tin oxide–coated slides, which were allowed to dry in a desiccator until matrix coating. Anesthesia-induced unconsciousness was confirmed before blood collection and animals were humanely euthanized before brain extraction according to the American Veterinary Medical Association Recommendations Guidelines for the Euthanasia of Animals: 2013 Edition.

Brain sections were manually sprayed with matrix (CHCA-acetonitrile/water/trifluoroacetic acid = 70/29.9/0.1) using an artistic airbrush (Procon Boy FWA Platinum 0.2-mm caliber airbrush, Mr. Hobby) to form the matrix crystal. MALDI tandem mass spectra were acquired on matrix-assisted laser desorption/ionization—time-of-flight (MALDI-TOF) (ultrafleXtreme, Bruker Daltonik GmbH) equipped with a Nd:YAG laser (24, 25). After scanning the sections, the laser spot areas were detected at a spot-to-spot center distance of 150 μm in each direction of the brain. The section surface was irradiated with the YAG laser in positive ion mode. MS data were reconstructed into MS images with a mass bin width of m/z ± 0.1 using FlexImaging 4.0 Software (Bruker Daltonik GmbH). ITMS data were depicted using a rainbow scale over a single range on a linear scale. The red and black pixels represent the highest (100%) and lowest (0%) signal intensity of a particular ion. Optical images of brain sections were obtained using a scanner (GT-X820, Canon), followed by MALDI-TOF-ITMS of the sections.

Data from the MTT assays and tumor progression of xenografts were expressed as means ± SD and as means ± SE, respectively. The sample size of in vivo experiments was shown in Supplementary Table S1. The comparison of two groups was analyzed by Student t test and more than two groups were analyzed by one-way ANOVA. P < 0.05, two-sided, was considered statistically significant. All statistical analyses were performed using GraphPad Prism Ver. 6 (GraphPad Software, Inc.).

A 70-year-old man presented with advanced papillary thyroid carcinoma, accompanied by multiple brain, bone, and pulmonary metastases. Following total thyroidectomy, he underwent WBI against the brain metastases. Because an iodine-131 (I-131) whole-body scan did not detect the accumulation of I-131 in the lesions, the tumor was diagnosed as refractory to I-131 therapy. One month after the completion of WBI, the patient was administered 16 mg/day lenvatinib, which reduced tumor size and induced the regression of edema around tumors (Fig. 1A). The patient's brain metastases were controlled for longer than 4 months. This favorable clinical course of this patient led to an evaluated lenvatinib efficacy against brain metastases of ATC in preclinical experiments.

In the first set of in vitro experiments, the effects of lenvatinib and sorafenib on cell viability was assessed using six ATC cell lines, OCUT-1C, OCUT-1F, OCUT-2, OCUT-4, OCUT-5, and OCUT-6. In agreement with previous findings (19), we found that neither lenvatinib nor sorafenib, at concentrations up to 1 μmol/L, affected the viability of these six cell lines, both in flat plate- and 3D-culture conditions (Fig. 1B; Supplementary Fig. S1). Because both lenvatinib and sorafenib inhibit angiogenesis-related receptors, we tested the effect of these agents on VEGFR2, PDGFR, and FGFR1 expression in ATC cells. Western blot analysis showed that all six ATC cell lines expressed total and phosphorylated PDGFRα and FGFR1, but not VEGFR2, at varying levels (Fig. 1C). Assays of cell production of angiogenesis-related cytokines showed that all six ATC cell lines secreted VEGF, VEGF-C, and angiopoietin-2 (ANG-2), but not basic fibroblast growth factor (Supplementary Table S2).

Because OCUT-1C and OCUT-2 produced subcutaneous tumors in SCID mice, we examined the effects of lenvatinib and sorafenib on the expression by these cells of angiogenesis-related molecules. Interestingly, at concentration of 1 μmol/L, which did not affect cell viability, lenvatinib and sorafenib inhibited VEGF production in OCUT-1C and OCUT-2 cells (Supplementary Table S3). Sorafenib also inhibited VEGF-C production in OCUT-1C cells. Moreover, 0.1 μmol/L lenvatinib and sorafenib markedly inhibited the phosphorylation of PDGFRα, FGFR1, and FRS2α (a downstream molecule of FGFR1) in vitro, but had no effect on the expression of cyclin A and cleaved-PARP (Fig. 1D). These results suggest that, while neither lenvatinib nor sorafenib directly affected cell viability, both could inhibit angiogenesis by suppressing the secretion of VEGF from ATC cell lines.

The effects of lenvatinib and sorafenib were compared in vivo utilizing the subcutaneous and brain metastasis–mimicking models. OCUT-1C and OCUT-2 cells developed into subcutaneous tumors in SHO mice, with oral administration of lenvatinib or sorafenib started after a tumor volume reached approximately 100 mm³. Compared with untreated mice, treatment with lenvatinib (10 mg/kg or 50 mg/kg) or sorafenib (25 mg/kg or 50 mg/kg) markedly inhibited the growth of tumors induced by OCUT-1C and OCUT-2 cells (Fig. 2). Treatment with 10 mg/kg lenvatinib and 50 mg/kg sorafenib showed similar growth inhibitory effects in these subcutaneous tumor models. None of the groups of mice showed significant weight loss (Supplementary Fig. S2A).

To elucidate the mechanisms underlying the antitumor effects of lenvatinib and sorafenib in vivo, tumor cell proliferation and microvessel density were evaluated by IHC staining of subcutaneous tumors. Treatment with both lenvatinib and sorafenib significantly reduced the number of Ki-67–positive proliferating tumor cells and microvessel densities as compared with the control group, (Fig. 3A–C). We further examined the expression of cell cycle and apoptosis markers in the subcutaneous tumors using Western blot analyses. While the expression of cyclin A was inhibited by treatment with lenvatinib or sorafenib, the expression of cleaved PARP was not altered (Supplementary Fig. S3). These results suggest that both lenvatinib and sorafenib inhibited the growth of subcutaneous tumors produced by ATC cells by inhibiting angiogenesis and by inducing cell-cycle arrest, but not by inducing tumor cell apoptosis.

We next examined the efficacy of lenvatinib and sorafenib in a brain metastasis–mimicking model. Only one of the six ATC lines tested in this study, OCUT-1C, induced brain lesions when inoculated into the brain parenchyma of SHO mice. Therefore, we chose OCUT-1C cells and established the EGFP-Eluc–transfected OCUT-1C/Eluc cell line for in vivo imaging (22). OCUT-1C and OCUT-1C/Eluc cells showed similar growth rates and sensitivities to lenvatinib and sorafenib in vitro (Supplementary Fig. S4A and S4B). Following the intracerebral inoculation of OCUT-1C/Eluc cells into SHO mice, bioluminescence was detected in the brain as early as day 10. These mice were subsequently randomized into three groups, with one group each treated with vehicle, lenvatinib (10 mg/kg), or sorafenib (50 mg/kg); these concentrations were chosen because they did not differ significantly in antitumor effects in a subcutaneous tumor model at day 21 or 27 after treatment initiation, respectively (P > 0.05; Fig. 2). Treatment with 10 mg/kg lenvatinib markedly reduced the growth of brain metastases. In contrast, unlike in the subcutaneous tumor model, treatment with 50 mg/kg sorafenib failed to inhibit the growth of brain tumors induced by OCUT-1C/Eluc cells (Fig. 4A and B; Supplementary Fig. S4C). Treatment with either lenvatinib or sorafenib had minimal effects on mouse body weights (Supplementary Fig. S2B).

We next assessed tumor cell proliferation and microvessel density by IHC staining of brain tumors after treatment with lenvatinib or sorafenib for 34 days. Compared with vehicle or sorafenib, lenvatinib treatment significantly reduced the number of Ki-67–positive proliferating tumor cells and CD31-positive endothelial cells in brain tumors (Fig. 4C–E). These results indicate that lenvatinib, but not sorafenib, inhibited tumor growth in the brain, presumably by inhibiting angiogenesis. In addition, we treated the mice with control, lenvatinib (10 mg/kg), or sorafenib (50 mg/kg) for 3 days (N = 3/group). When we evaluated angiogenesis by measuring immunofluorescence for CD31 in brain and subcutaneous tumors, we found that treatment with lenvatinib for 3 days markedly decreased angiogenesis in subcutaneous tumors and brain, whereas treatment with sorafenib for 3 days reduced angiogenesis in the subcutaneous tumors but not in brain tumors. These results indicated that lenvatinib, but not sorafenib, rapidly inhibited angiogenesis in brain tumor lesions (Fig. 4F; Supplementary Fig. S5A–S5C).

Finally, we evaluated the distribution of lenvatinib and sorafenib in the tumors of the brain and subcutaneous space using the ITMS method with results corrected at m/z 427.4 and m/z 465.8, representing the precursor ions of lenvatinib or sorafenib, respectively (Supplementary Fig. S6A and S6B). Laser power was optimized to minimize in-source decay of anticancer drugs. The obtained precursor ions were cleaved at the cyclopropylamino- or carbamoylamino-moieties to provide ions with m/z 229.1 and 370.1, respectively.

Lenvatinib was distributed to tumors and nontumor parenchyma in both the brain and the subcutaneous space. Sorafenib was also distributed to tumors and nontumor parenchyma in the subcutaneous space. In the brain, however, slight amounts of sorafenib were detected in tumors but none was present in the nontumor parenchyma (Fig. 5A). The peak intensity values of the spectra were normalized by dividing them by the total ion current for semiquantitative comparisons of tumor and nontumor regions. These values were expressed as the coefficient of variation of mean + CV (P < 0.05 by Student t test). The relative amounts of lenvatinib were significantly higher in tumor than in nontumor regions of brain (7.6 ± 0.1 vs. 4.7 ± 0.1; P < 0.01) and skin (3.4 ± 0.5 vs. 1.0 ± 0.2; P < 0.05; Fig. 5B). In contrast, the relative amounts of sorafenib at tumor and nontumor regions were similar in brain (1.7 ± 0.1 vs. 1.3 ± 0.3) and skin (0.8 ± 0.1 vs. 1.0 ± 0.2; Fig. 5B).

These results clearly indicate that, although both lenvatinib and sorafenib penetrate subcutaneous tumors, only lenvatinib penetrates brain tumors, a finding consistent with the efficacy of these two agents against subcutaneous and brain tumors.

This study showed that the multikinase inhibitor lenvatinib could inhibit angiogenesis and thereby suppress the growth of ATC cells not only in the subcutaneous space but also in the brain. Conversely, sorafenib inhibited the growth of ATC cells in the subcutaneous space but not in the brain. The lower efficacy of sorafenib against brain lesions was due to its limited penetration into the brain. These findings suggest that lenvatinib may be useful for controlling brain metastases in patients with ATC.

We previously reported that VEGF is essential for the production of brain metastases via the promotion of angiogenesis and that inhibition of the VEGF/VEGFR axis could control the progression of brain metastasis by suppressing angiogenesis (26). Many other studies have showed that VEGF plays a pivotal role in the subcutaneous growth of various types of tumor and that VEGF inhibition markedly suppresses tumor growth (27–29). Both lenvatinib and sorafenib suppress angiogenesis by inhibiting VEGFR-2 (30, 31). In agreement with these previous findings, we observed that: (i) neither lenvatinib nor sorafenib directly inhibited the growth of ATC cells in vitro; (ii) both lenvatinib and sorafenib inhibited VEGF production by ATC cells in vitro; and (iii) both lenvatinib and sorafenib markedly reduced tumor vessel density and the number of proliferating tumor cells, except that sorafenib did not affect these parameters in the brain metastasis model. These findings collectively suggest that the antitumor effects of lenvatinib and sorafenib in these brain metastasis and subcutaneous tumor models with ATC cells are predominantly due to their inhibition of angiogenesis. Similarly, lenvatinib was associated with antitumor activity in advanced glioblastoma (32). Biomarkers predicting the efficacy of antiangiogenetic agents include p-VEGFR2 expression in advanced breast cancer and angiopoietin 2 expression in differentiated cancer of the thyroid (33, 34). Further investigations are warranted to determine which tumors respond to antiangiogenesis inhibitors, including lenvatinib.

Interestingly, while both lenvatinib and sorafenib inhibited the growth of ATC cells in the subcutaneous space, only lenvatinib inhibited the growth of ATC in the brain. The major difference between these two compounds associated with differences in efficacy in the brain metastasis model was in their distribution within brain tumors. Using the ITMS method to visualize the distribution of these targeted agents in brain tumor lesions, we found that the distribution of sorafenib in the brain parenchyma was much lower than its distribution in the subcutaneous space, suggesting that sorafenib may be unable to penetrate the blood–brain barrier (BBB). The finding of slight amounts of sorafenib in brain tumor lesions suggests that tumor cells may partially disrupt the BBB (35). Conversely, lenvatinib was distributed throughout the tumor lesions and nontumor parenchyma of the brain and subcutaneous space. Because both lenvatinib and sorafenib have been reported to be substrates of P-glycoprotein, an efflux pump on endothelial cells that comprise the BBB, penetration (influx) into the brain may differ for these two drugs. Intensive assessment of targeted drug distribution using the ITMS method combined with histologic analyses may lead to the future development of more effective drugs against brain metastases.

WBI is known as one of the most useful tools for controlling brain metastases, however its efficacy against brain metastases of ATC remains unclear. While our patient had brain metastases of papillary thyroid carcinoma, his brain metastases were well controlled by lenvatinib monotherapy for longer than 4 months, consistent with the results of our in vivo experiments. Combined use of WBI and lenvatinib might be more effective to brain metastasis of ATC. Careful clinical evaluation for such combined therapy is warranted in future.

A recent clinical trial demonstrated that combined treatment with the MEK inhibitor trametinib and the BRAF inhibitor dabrafenib was effective in patients with BRAF^V600-mutated ATC. The overall response rate was 69%, and Kaplan–Meier estimates at 12 months of duration of response, progression-free survival, and overall survival rates were 90%, 79%, and 80%, respectively (36), but outcomes were not reported in patients with brain metastases. Generally, most patients who experience an initial response relapse due to acquired resistance (37, 38). Moreover, the incidence of brain metastasis, which is life-threatening, should increase in association with prolonged survival resulting from improved cancer treatment (39, 40). Therefore, a novel therapy to overcome the resistance of brain metastases is necessary for further improvement of patient survival. The ATC cell line OCUT-1C, which possesses the BRAF^V600E mutation, is sensitive to trametinib and dabrafenib in vitro (41). It will be of interest to assess the effect of combined treatment with trametinib and dabrafenib in the brain metastasis–mimicking model with OCUT-1C cells. Moreover, the efficacy of lenvatinib should be evaluated in models of brain metastases with acquired resistance to combined treatment with trametinib and dabrafenib.

The major limitation of this study was our ability to establish a brain tumor model with only one of the six ATC cell lines, OCUT-1C cells. Further development of mouse models is warranted to evaluate drug efficacy and the delivery of targeted agents to ATC cells.

In conclusion, this study demonstrated that lenvatinib has activity against brain lesions produced by human ATC cells. These findings provide a rationale for the lenvatinib treatment of patients with ATC with brain metastases.

N. Onoda reports receiving commercial research grant from Bayer, has received speakers bureau honoraria from Eisai, Bayer, and Sanofi, and is a consultant/advisory board member for Bayer. S. Yano report receiving commercial research grant from Eisai. No potential conflicts of interest were disclosed by the other authors.

Conception and design: T. Yamada, J. Matsui, S. Yano

Development of methodology: R. Wang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Wang, S. Arai, K. Fukuda, H. Taniguchi, S. Takeuchi, K. Yamashita, E. Hirata, K. Ohtsubo

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Wang, T. Yamada, A. Tanimoto, S. Taira

Writing, review, and/or revision of the manuscript: R. Wang, T. Yamada, J. Matsui, N. Onoda, S. Yano

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Wang, A. Tanimoto, A. Nishiyama, J. Matsui, N. Onoda, S. Yano

Study supervision: T. Yamada, S. Yano

We thank Yukina Tatsuta, Department of Bioscience, Fukui Prefectural University, for assisting with sample preparation for ITMS. This work was supported by a research grant from Eisai Co. Ltd. (to S. Yano), JSPS KAKENHI grant no. 16H05308 (to S. Yano), the Project for Cancer Research and Therapeutic Evolution (P-CREATE) grant no. 16cm0106513h0001 (to S. Yano), Research Grant of the Princess Takamatsu Cancer Research Fund grant no. 16-24822 (to S. Yano), and for developing innovative cancer chemotherapy from the Kobayashi Foundation for Cancer Research (to T. Yamada).

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.