Foretinib Overcomes Entrectinib Resistance Associated with the NTRK1

Author Manuscript Published OnlineFirst on February 20, 2018; DOI: 10.1158/1078-0432.CCR-17-1623 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

1 CCR

2 Foretinib overcomes entrectinib resistance associated with the NTRK1

3 G667C mutation in NTRK1 fusion-positive tumor cells in a brain

4 metastasis model

6 Akihiro Nishiyama1, Tadaaki Yamada1,2, Kenji Kita1, Rong Wang1, Sachiko Arai1,

7 Koji Fukuda1, Azusa Tanimoto1, Shinji Takeuchi1, Shoichiro Tange3, Atsushi

8 Tajima3, Noritaka Furuya4, 5, Takayoshi Kinoshita4, Seiji Yano1, #

9 10 1Division of Medical Oncology, Cancer Research Institute, Kanazawa University, 11 Kanazawa, Japan 12 2Department of Pulmonary Medicine, Graduate School of Medical Science, Kyoto 13 Prefectural University of Medicine, Kyoto, Japan 14 3Department of Bioinformatics and Genomics, Graduate School of Advanced 15 Preventive Medical Sciences, Kanazawa University, Kanazawa, Japan 16 4Graduate School of Science, Osaka Prefecture University, Osaka, Japan 17 5Kissei Pharmaceutical, Nagano, Japan 18 19 20 # Corresponding Author: 21 Seiji Yano, MD, PhD. 22 Division of Medical Oncology, Cancer Research Institute, Kanazawa University 23 13-1, Takaramachi, Kanazawa, Ishikawa 920-0934, Japan 24 Phone: +81-76-265-2794, Fax: +81-76-244-2454 25 E-mail: [email protected] 26 27 Word count: 5752 28 Number of figures: 6; Number of tables: 0 29 Supporting information: 9 30 31 Conflict of interest: Dr. Tadaaki Yamada received research grant from Boehringer 32 Ingelheim. Other authors have no potential conflicts of interest.

Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2018 American Association for Cancer Research. Author Manuscript Published OnlineFirst on February 20, 2018; DOI: 10.1158/1078-0432.CCR-17-1623 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

1 2 Running title: Foretinib overcomes entrectinib resistance in brain tumor 3 4 Key words: NTRK1, entrectinib, foretinib, drug resistance, colon cancer 5

1 Translational relevance

2 The CNS is a sanctuary site from targeted drugs. However, the mechanism of targeted

3 drug resistance in CNS tumors is largely unknown. Entrectinib, an inhibitor of multiple

4 kinases such as ALK, ROS1, and TRK, is clinically effective in the treatment of cancer

5 patients with rearrangements in ALK, ROS1, and NTRK1. While entrectinib has initial

6 activity against CNS metastases, resistance usually develops. Here, we found that the

7 NTRK1-G667C mutation, which causes moderate resistance to entrectinib in vitro, could

8 be detected in a brain metastasis-mimicking model when acquired resistance to

9 entrectinib was induced. Furthermore, we demonstrated that foretinib could inhibit the

10 phosphorylation of TRK-A with the G667C mutation and overcome entrectinib

11 resistance in the animal models for liver and brain metastases. Our findings provide a

12 rationale for clinical trials with foretinib in cancer patients with entrectinib-resistant

13 tumors harboring the NTRK1-G667C mutation, including patients with brain metastases.

1 Abstract

2 Background

3 Rearrangement of the neurotrophic tropomyosin receptor kinase 1 (NTRK1) gene, which

4 encodes tyrosine receptor kinase A (TRK-A), occurs in various cancers, including colon

5 cancer. Although entrectinib is effective in the treatment of central nervous system

6 (CNS) metastases that express NTRK1 fusion proteins, acquired resistance inevitably

7 results in recurrence. CNS is a sanctuary for targeted drugs; however, the mechanism by

8 which CNS metastases become entrectinib-resistant remains elusive and must be

9 clarified to develop better therapeutics.

10 Experimental design

11 The entrectinib-resistant cell line KM12SM-ER was developed by continuous treatment

12 with entrectinib in the brain metastasis-mimicking model inoculated with the

13 entrectinib-sensitive human colon cancer cell line KM12SM, which harbors the

14 TPM3-NTRK1 gene fusion. The mechanism of entrectinib resistance in KM12SM-ER

15 cells was examined by next-generation sequencing. Compounds that overcame

16 entrectinib resistance were screened from a library of 122 kinase inhibitors.

17 Results

18 KM12SM-ER cells, which showed moderate resistance to entrectinib in vitro, had

19 acquired the G667C mutation in NTRK1. The kinase inhibitor foretinib inhibited TRK-A

20 phosphorylation and the viability of KM12SM-ER cells bearing the NTRK1-G667C

21 mutation in vitro. Moreover, foretinib markedly inhibited the progression of

22 entrectinib-refractory KM12SM-ER-derived liver metastases and brain tumors in

23 animal models, predominantly through inhibition of TRK-A phosphorylation.

24 Conclusion

1 These results suggest that foretinib may be effective in overcoming entrectinib

2 resistance associated with the NTRK1-G667C mutation in NTRK1 fusion-positive

3 tumors in various organs, including the brain, and provide a rationale for clinical trials

4 of foretinib in cancer patients with entrectinib-resistant tumors harboring the

5 NTRK1-G667C mutation, including patients with brain metastases.

1 Introduction

2 The neurotrophic tropomyosin receptor kinase 1 (NTRK1) gene, located at 1q21-q22,

3 encodes tyrosine receptor kinase A (TRK-A), which is a member of the TRK family (1).

4 The rearrangement of the NTRK1 locus is recognized as an oncogenic driver and occurs

5 across different tumors, including non-small cell lung cancer (NSCLC) and colon cancer,

6 though at a low incidence (~1%) (2). At least four NTRK1 fusion partners, such as

7 LMNA, TPM3, SQSTM1, and BCAN, have been reported (3-6). The NTRK1 fusion gene

8 products promote signal transduction through the PI3K-AKT, PLC-PKC, and

9 SHC-RAS-MAPK signaling pathways and play crucial roles in cell survival (7). Several

10 compounds, including crizotinib and entrectinib, have been shown to inhibit the growth

11 of tumor cells that express NTRK1 fusion proteins and have demonstrated remarkable

12 clinical response in patients with NTRK1 fusion-positive tumors (8).

13 Entrectinib is an orally available inhibitor of multiple tyrosine kinases, including

14 ALK, ROS1, and TRK-A/TRK-B/TRK-C (9). Entrectinib has been shown to be active

15 and well tolerated in NTRK1 fusion-positive tumor patients, including patients with

16 central nervous system (CNS) metastases (10), and is currently being developed in a

17 phase II clinical trial (NCT02568267). However, the response to entrectinib is limited in

18 time due to acquired resistance. Recent reports have demonstrated that the

19 NTRK1-G595R and G667C mutations drive acquired resistance to entrectinib in colon

20 cancer with NTRK1 fusion products (11). No effective drug that overcomes the

21 resistance associated with the NTRK1-G595R and G667C mutations has been developed

22 yet.

23 Targeted drugs, such as EGFR tyrosine kinase inhibitors (EGFR-TKIs) and

24 ALK-TKIs, are generally effective in the treatment of patients with CNS diseases, such

1 as brain metastasis and leptomeningeal carcinomatosis, with no history of TKI

2 administration (TKI-naïve). However, patients frequently experience acquired resistance

3 to targeted drugs during the progression of CNS metastases (12). The CNS is therefore

4 called “the sanctuary site of targeted drugs” (13). The mechanisms of targeted drug

5 resistance in brain lesions may be different from those in extra-cranial lesions. For

6 instance, the EGFR-T790M mutation is responsible for 50–60% of the acquired

7 resistance in extra-cranial lesions from EGFR-mutated NSCLC treated with

8 EGFR-TKIs, while it is much less frequent in CNS lesions from the same patients (14).

9 While poor penetration of targeted drugs into the CNS, because of the blood–brain

10 barrier (BBB), is thought to be associated with the drug resistance (15), the mechanism

11 of resistance in CNS lesions, whose understanding is important to establish better

12 therapeutics, remains elusive.

13 We previously reported (16) that KM12SM cells, a highly metastatic variant of

14 the KM12C colon cancer cell line, which have been shown to express the

15 TPM3-NTRK1 fusion protein (17), have a high potential to produce brain metastases

16 after internal carotid artery inoculation in nude mice. The incidence of brain metastases

17 in colon cancer, previously considered rare, has now increased and is associated with

18 increased survival owing to improved treatments (18). Based on these considerations,

19 we examined the mechanism of entrectinib resistance in brain lesions, utilizing a brain

20 metastasis-mimicking model obtained through the injection of KM12SM cells. We

21 found that the NTRK1-G667C mutation caused entrectinib resistance in brain lesions.

22 Furthermore, we demonstrated, for the first time, that foretinib can overcome entrectinib

23 resistance associated with the NTRK1-G667C mutation in the liver metastasis model

24 and the brain metastasis-mimicking model.

3 Materials and methods

4 Cell lines, cell culture, and compounds

5 The human colon cancer cell line KM12SM was kindly gifted by Dr. I. J. Fidler (M.D.

6 Anderson Cancer Center, Houston, TX) to our group in 1999. KM12SM cells are a

7 highly metastatic variant of KM12C cells, which present the TPM3-NTRK1 gene

8 rearrangement (17). The KM12SM cell line (which generates spontaneous metastasis)

9 was established from liver metastases after orthotopic cecum inoculation of cultured

10 KM12C cells in nude mice (19). KM12SM cells were maintained and cultured in

11 RPMI-1640 medium (Thermo Fischer Scientific K.K., Kanagawa, Japan) with 10%

12 fetal bovine serum (FBS), penicillin (100 units/mL), and streptomycin (50 µg/mL), in a

13 humidified CO2 incubator at 37 C. Ba/F3, Ba/F3 transfected with wild-type

14 TPM3-NTRK1 (Ba/F3_WT), Ba/F3 transfected with TPM3-NTRK1 G595R

15 (Ba/F3_G595R), and Ba/F3 transfected with TPM3-NTRK1 G667C (Ba/F3_G667C)

16 cells were gifted by Dr. R. Katayama (Division of Experimental Chemotherapy, Cancer

17 Chemotherapy Center, Japanese Foundation for Cancer Research, Tokyo, Japan). Ba/F3

18 cells were cultured in DMEM containing 10% FBS with 0.5 ng/ml interleukin (IL)-3

19 (Invitrogen). Ba/F3_WT, Ba/F3_G595R, and Ba/F3_G667C cells were cultured in

20 DMEM containing 10% FBS. Possible mycoplasma infection in the cells was regularly

21 checked using a MycoAlert Mycoplasma Detection Kit (Lonza, Basel, Switzerland).

22 Cell line authentication was performed by short tandem repeat analysis in a laboratory

23 at the National Institute of Biomedical Innovation (Osaka, Japan) in May 2015. Human

24 dermal microvascular endothelial cells (HMVEC) were purchased from KURABO

1 (Osaka, Japan), maintained in HuMedia-MvG with growth supplements (KURABO),

2 and used for in vitro assays at passages two to five. Entrectinib, foretinib, and

3 nintedanib were purchased from Selleck Chemicals (Houston, TX, USA).

4 Cell viability assay

5 Cells were seeded at a density of 2 × 103/100 μL RPMI-1640 plus 10% FBS per well

6 in 96-well culture plates and incubated overnight. The next day, the indicated compound

7 was added to each well at a concentration gradient (0, 0.003, 0.01, 0.1, 0.3, 1, 3, 10, and

8 30 nmol/L), and incubation was continued for an additional 72 h. Cell viability was

9 measured using a CCK-8 kit (Dojindo Laboratories, Kumamoto, Japan). Each

10 experiment was performed at least three times, with triplicate samples. Ba/F3,

11 Ba/F3_WT, Ba/F3_G595R, and Ba/F3_G667C cells were seeded at a density of 3 ×

12 103/100 μL DMEM plus 10% FBS with or without IL-3 per well in 96-well culture

13 plates and incubated overnight. The following procedures were the same.

14 Antibodies and western blotting

15 Cells were lysed with Cell Lysis Buffer (Cell Signaling Technology, Beverly, MA,

16 USA) containing 1% (v/v) phosphatase inhibitor cocktail 3 (Sigma, St Louis, IL, USA)

17 and distilled water. Cell extracts (20 μg per lane) were separated by SDS-PAGE using

18 Mini-PROTEAN TGX Precast Gels, and the separated proteins were electrophoretically

19 transferred to Immun-Blot PVDF membranes (Bio-Rad). The primary antibodies used

20 were: TRKA (14G6), phospho-TRKA (Tyr490), AKT, phospho-AKT (Ser473), MET

21 (25H2), VEGFR2 (55B11), cleaved PARP (Asp214), β-actin (all from Cell Signaling

22 Technology), ERK1/ERK, phospho-ERK1/ERK2 (Thr202/Thr204) (both from R&D

23 System, Minneapolis, MN, USA), and TRKA (C-14) (Santa Cruz Biotechnology, Dallas,

24 TX, USA). The membranes treated with the primary antibodies were incubated for 1 h

1 at temperature range with species-specific HRP-conjugated secondary antibodies.

2 Immunoreactive bands were visualized with SuperSignal West Dura Extend Duration

3 Substrate, an enhanced chemiluminescent substrate (Pierce Biotechnology, Rockford, IL,

4 USA).

5 RNA sequencing

6 Total RNA was extracted from tumor cells using an RNeasy® Plus Mini kit (Qiagen,

7 Valencia, CA, USA). Quantity and quality of the extracted RNAs were evaluated with

8 the Qubit RNA BR Assay kit (Thermo Fisher Scientific, Carlsbad, CA, USA) and the

9 RNA 6000 Nano LabChip on the Agilent 2100 Bioanalyzer (Agilent Technologies,

10 Santa Clara, CA, USA). One microgram of total RNA was used for the construction of a

11 strand-specific RNA sequencing library. The library was prepared using the TruSeq

12 Stranded mRNA LT Sample Prep kit (Illumina, San Diego, CA, USA) according to the

13 manufacturer’s instructions. The individually indexed libraries were pooled, and then

14 diluted to 10 pM for cluster formation with the TruSeq PE Cluster kit v3-cBot-HS

15 (Illumina), using an Illumina cBot fluidics station. RNA sequencing analysis was

16 performed using a HiSeq2000 sequencer (Illumina) with a 2  100 bp paired-end

17 module. The alignment of the paired-end FASTQ files to the human reference genome

18 (GRCh37) was performed using TopHat2 software (20), and visualized using Integrative

19 Genomics Viewer (IGV) software (21). The FASTQ files were further analyzed, to find

20 known and/or novel fusion transcripts, using FusionCatcher software (22). To validate

21 the presence of the TPM3-NTRK1 fusion transcript and a mutation in the kinase domain,

22 the total RNA extracted from the two cell lines was also subjected to reverse

23 transcription-PCR (RT-PCR) and Sanger sequencing analysis. In brief, complementary

24 DNAs (cDNAs) for PCR templates were synthesized from the total RNA with the

1 SuperScript VILO cDNA Synthesis kit (Thermo Fisher Scientific). DNA fragments

2 encompassing both the TPM3-NTRK1 fusion break-point and the mutation site were

3 generated by PCR amplification using TaKaRa Ex Taq DNA polymerase (TAKARA

4 BIO, Shiga, Japan). Nucleotide sequences of the fragments were determined with the

5 BigDye Terminator v3.1 Cycle Sequencing kit (Thermo Fisher Scientific) on an Applied

6 Biosystems 3130 Genetic Analyzer (Thermo Fisher Scientific).

7 RNA interference

8 Duplexed Silencer Select siRNAs against NTRK1 (s9745, s9746), Silencer Select

9 siRNAs against VEGFR2 (s7822, s7833), siRNAs targeting MET (HSS106478), and

10 Silencer® Select Negative Control #1 siRNA (Ambion®) were used for RNA

11 interference (RNAi) assays. Briefly, aliquots of 1 × 105 cells in 2 mL of antibiotic-free

12 medium were plated into each well of a 6-well plate and incubated at 37 °C for 24 h.

13 The cells were transfected with siRNA (250 pmol) or scrambled RNA using

14 Lipofectamine RNAiMAX (5 μL, Invitrogen) in accordance with the manufacturer’s

15 instructions. After 24 h, the cells were washed twice with PBS and incubated for an

16 additional 48 h in antibiotic-containing medium. The tumor cells were then used for cell

17 proliferation assays. NTRK1, VEGFR2, and MET knockdowns were confirmed by

18 western blot.

19 Kinase inhibitor library screen

20 We used the Kinase Inhibitor Library (L1200-01 and -02) from Selleck Chemicals for

21 the screening of kinase inhibitors. Briefly, cancer cells were seeded (2,000 cells/well)

22 overnight into 96-well plates and incubated with 0.1 μM and 1 μM of 122 kinase

23 inhibitors for 72 h. The inhibitory activity of each drug was assessed with a CCK-8 kit.

24 EGFP-Eluc gene transfection

1 The expression vector GFP-Eluc (Enhanced green fluorescent protein [EGFP] from the

2 pIRES-EGFP Vector and Emerald Luc [Eluc] from Emerald Luc Vector) (23) was

3 transfected into KM12SM cells using Lipofectamine LTX (Invitrogen) according to the

4 manufacturer’s instructions. The cells were then selected with puromycin.

5 Docking analyses

6 The coordinates of TRK-A for the docking study of foretinib and entrectinib were

7 prepared from the X-ray structures with DFG-out conformation (PDB code: 5JFS) and

8 DFG-in conformation (PDB code: 4AOJ), respectively. The coordinate of the G667C

9 mutant for the docking study of foretinib was prepared using the Discovery Studio

10 (version 4.5, BIOVIA, San Diego, CA, USA). The three-dimensional conformations of

11 compounds were generated by OMEGA (version 3.141592-1.23.2.3, OpenEye

12 Scientific Software, Santa Fe, NM, USA) with default settings. Docking was performed

13 using FRED (version 3.0.1, OpenEye Scientific Software) with constraints making a

14 hydrogen bond with Met592. The displayed docking models had the best FRED score.

15 Animal experiments

16 All animal experiments in this study were performed in strict accordance with the

17 recommendations in the Guide for the Care and Use of Laboratory Animals of the

18 Ministry of Education, Culture, Sports, Science and Technology, Japan. The protocol

19 was approved by the Committee of Ethics of Experimental Animals, and the Advanced

20 Science Research Center, Kanazawa University, Kanazawa, Japan (Approval number:

21 AP-081088). Five-week-old female SCID mice (C.B-17/Icr-scid/scidJc) and male

22 Hairless SCID (SHO®) mice were purchased from CLEA (Tokyo, Japan) and Charles

23 River Laboratories Japan (Yokohama, Japan), respectively.

24 For the liver metastasis model, tumor cells were inoculated into the spleen as reported

1 previously (24). In brief, the SCID mice were anesthetized with tribromoethanol and

2 inhalation of 1.5% isoflurane. After an incision was made in the skin, subcutaneous

3 tissue, and peritoneal membrane on the left side of the abdomen, the spleen was

4 exposed. A suspension of tumor cells (1 × 106/50 μL) was injected into the spleen

5 using a 27-G needle. The incision was closed by suture.

6 For the brain metastasis-mimicking model, tumor cells were inoculated into the brain

7 as reported previously (23). In brief, the SHO mice were anesthetized with

8 tribromoethanol. The scalp was sterilized with 70% ethanol, and a small hole was

9 created in the skull, 0.5 mm anterior and 3.0 mm lateral to the bregma, using a dental

10 drill. A suspension of tumor cells (1.5 × 105/1.5 μL) was injected into the right

11 striatum, 3 mm below the surface of the brain, using a 10-μL Hamilton syringe with a

12 26-G needle. The scalp was closed using an Autoclip Applier (BD, Franklin Lakes, NJ,

13 USA).

14 In both models, the development of tumors was tracked in live mice by repeated

15 non-invasive optical imaging of tumor-specific luciferase activity using the IVIS

16 Lumina XR Imaging System (PerkinElmer, Alameda, CA) as described previously (25).

17 The intensity of the bioluminescence signal was analyzed using Living Image 4.0

18 software (PerkinElmer) by serially quantifying the peak photon flux in a selected region

19 of interest (ROI) within a given tumor. The intensity of the bioluminescence signal was

20 corrected to consider the total area of the ROI and the elapsed time for which

21 bioluminescence signals were read with a CCD camera. The corrected value was then

22 expressed as photons/s.

23 Liver: Tumor luminescence and mouse body weight were measured twice a week.

24 Thirteen days after inoculation, mice were orally administered vehicle, entrectinib (15

1 mg/kg), or foretinib (30 mg/kg) daily, for 10 consecutive days.

2 Brain: After seven days, the mice were randomized and treated orally with saline,

3 entrectinib (15 mg/kg), or foretinib (30 mg/kg), daily. The mouse body weights were

4 measured twice per week.

5 Immunohistochemistry (IHC)

6 Formalin-fixed, paraffin-embedded tissue sections (4 μm thick) were de-paraffinized.

7 Proliferating cells were detected by incubating tissue sections with a Ki-67 antibody

8 (Clone MIB-1; DAKO Corp, Glostrup, Denmark). Apoptotic cells were detected by

9 TUNEL staining (In Situ Cell Death Detection Kit; Roche, Mannheim, Germany) of the

10 tissue sections. Antigens were retrieved by microwaving the tissue sections in 10 mM

11 citrate buffer (pH 6.0) and Tris-EDTA buffer (pH 9.0). After incubation with a

12 secondary antibody, peroxidase activity was visualized via a DAB reaction using a

13 Histofine Simple Stain MAX-PO(R) kit (Nichirei, Tokyo, Japan). The sections were

14 counterstained with hematoxylin. All sections were also stained with hematoxylin and

15 eosin.

16 Quantification of immunohistochemistry results

17 The five areas containing the highest numbers of stained cells within each section were

18 selected for histological quantification, via light or fluorescence microscopy at 400×

19 magnification. All results were independently evaluated by two authors (T.Y. and A.N.).

20 Statistical analysis

21 Data from the viability assays and the tumor progression in the xenograft model are

22 expressed as mean ± S.D. The statistical significance of differences was analyzed by

23 one-way ANOVA and Spearman rank correlations performed using GraphPad Prism Ver.

24 6.0 (GraphPad Software, Inc., San Diego, CA, USA). For all analyses, a two-sided P

1 value less than 0.05 was considered statistically significant.

1 Results

2 Establishment of KM12SM-ER cells that acquired in vivo resistance to entrectinib.

3 We previously confirmed that both the KM12C and KM12SM cells had the

4 TPM3-NTRK1 gene rearrangement and were equally sensitive to TRK inhibitors, such

5 as crizotinib and entrectinib (unpublished data). In this study, we first established

6 KM12SM/Eluc cells by transfecting a plasmid expressing the EGFP/Eluc gene into

7 KM12SM cells. KM12SM and KM12SM/Eluc cells were similarly sensitive to

8 entrectinib in vitro (Supplementary Fig. 1), indicating that the transfection of

9 EGFP-Eluc had no effect on the sensitivity to entrectinib. To establish resistant tumor

10 cells that had acquired entrectinib resistance in the microenvironment of the brain,

11 KM12SM/Eluc cells were inoculated into the brain of SCID mice. The treatment with

12 entrectinib (15 mg/kg), commenced seven days after tumor cell inoculation, discernibly

13 delayed the progression of tumors in the brain and prolonged the survival of the mice,

14 indicating the antitumor effect of entrectinib on KM12SM/Eluc cells in the brain.

15 However, the tumor luminescence gradually increased, despite the continued entrectinib

16 treatment, suggesting that the tumor cells in the brain had acquired resistance to

17 entrectinib. We harvested the brain tumor from mice on day 45 (entrectinib treatment

18 was given for 37 days) and cultured it in vitro. The expanded tumor cells were named

19 KM12SM-ER (Fig. 1A). We found that the KM12SM-ER cells were resistant to

20 treatment with 30 mg/kg entrectinib in our brain metastasis-mimicking model

21 (Supplementary Fig. 2).

22 Next, we examined the sensitivity of KM12SM-ER cells in vitro. As expected,

23 KM12SM-ER cells were discernibly resistant to entrectinib compared with KM12SM

24 cells (Fig. 1B; IC50 KM12SM-ER, 56.1 nM; IC50 KM12SM, 1.70 nM). To assess the

1 resistance mechanism, we examined the effect of entrectinib on the phosphorylation of

2 TRK-A and its downstream molecules by western blot. Entrectinib inhibited

3 phosphorylation of TRK-A and its downstream molecules AKT and ERK, and thereby

4 induced apoptosis, indicated by the expression of cleaved PARP (c-PARP) in KM12SM

5 cells, but not in KM12SM-ER cells (Fig. 1C). These results suggest that entrectinib

6 failed to inhibit TRK phosphorylation in KM12SM-ER cells, which were therefore

7 resistant to the drug. We next determined the effect of TRK-A knockdown in

8 KM12SM-ER cells utilizing siRNAs specific for NTRK1. Three different siRNAs for

9 NTRK1 successfully knocked down TRK-A expression and reduced the viability of both

10 KM12SM and KM12SM-ER cells (Fig. 1D and E). These results strongly suggest that

11 KM12SM-ER cells acquired entrectinib resistance through a TRK-A-dependent

12 mechanism, possibly through the acquisition of resistance-inducing mutations in

13 NTRK1.

14 KM12SM-ER cells had the NTRK1-G667C mutation.

15 We next performed next-generation sequencing using the RNA extracted from the

16 KM12SM/Eluc and KM12SM-ER cells. The sequencing analysis yielded 29,876,111

17 and 28,133,224 paired-end reads for the KM12SM/Eluc and KM12SM-ER cells,

18 respectively. Both the KM12SM/Eluc and KM12SM-ER cells possessed the

19 TPM3-NTRK1 fusion gene (Fig. 2A), whose presence was further confirmed by direct

20 sequencing of cDNA (Fig. 2C) and RT-PCR (Fig. 2B). Importantly, we found the

21 mutation G667C in exon 10 of the NTRK1 gene from the KM12SM-ER cells; this

22 mutation was not present in the KM12SM/Eluc cells. A recent study demonstrated two

23 resistance-inducing mutations, G595R and G667C, in the NTRK1 gene, through the

24 analysis of entrectinib-resistant KM12 cells and clinical specimens from liver

1 metastases and plasma DNA of patients with colon cancer who had acquired entrectinib

2 resistance (11). The G595R and G667C mutations were reported to induce high and

3 moderate resistance to entrectinib, respectively (11). Therefore, we concluded that

4 KM12SM-ER cells acquired entrectinib resistance through the NTRK1-G667C

5 mutation.

6 Foretinib and nintedanib can inhibit TRK-A bearing a G667C mutation.

7 Because no effective drug has been identified, so far, as being capable of inhibiting

8 TRK-A with a G667C mutation, we next sought to screen compounds which might have

9 such activity using a Kinase Inhibitor Library that includes 122 inhibitors (Fig. 3A).

10 Cell viability assays on KM12SM-ER cells revealed that some inhibitors could inhibit

11 the viability of KM12SM-ER cells. Specifically, 13 compounds inhibited the viability

12 by 40% compared with the control. Of the 13 compounds, five, including foretinib and

13 nintedanib, are known to inhibit VEGFR-2 (Fig. 3B). Because foretinib and nintedanib

14 are currently used in clinical trials or clinical practice, we further assessed the effect of

15 these two drugs.

16 As expected, foretinib (IC50 = 1.95 nmol/L) and nintedanib (IC50 = 27.5 nmol/L), but

17 not entrectinib (IC50 = 56.1 nmol/L), inhibited the viability of KM12SM-ER cells at a

18 concentration lower than 1 nmol/L (Fig. 3C). In addition to entrectinib, foretinib and

19 nintedanib also inhibited the viability of KM12SM cells. Interestingly, KM12SM-ER

20 cells were more sensitive to foretinib and nintedanib, compared with KM12SM cells.

21 The kinase inhibition profile assay also showed that both foretinib and nintedanib were

22 able to inhibit TRK-A, although their activity was weaker than that of entrectinib. The

23 activity of foretinib and nintedanib against TRK-B and TRK-C was much weaker than

24 that against TRK-A (Supplementary Fig. 3). Western blot analyses revealed that

1 entrectinib, foretinib, and nintedanib inhibited the phosphorylation of TRK-A and its

2 downstream molecules, AKT and ERK, in KM12SM cells, and entrectinib was the most

3 potent inhibitor. Foretinib and nintedanib inhibited the phosphorylation of TRK-A with

4 the G667C mutation in KM12SM-ER cells more efficiently than that of the wild type

5 TRK-A in KM12SM cells (Supplementary Fig. 4). Interestingly, foretinib and

6 nintedanib remarkably inhibited the phosphorylation of TRK-A, AKT, and ERK in

7 KM12SM-ER cells (Fig. 3D), whereas entrectinib mildly suppressed the

8 phosphorylation of TRK-A with the G667C mutation and transiently inhibited the

9 phosphorylation of ERK. (Supplementary Fig. 5). With regard to cell viability and

10 TRK-A phosphorylation, entrectinib induced apoptosis (assessed by expression of

11 c-PARP) in KM12SM cells, and foretinib and nintedanib caused apoptosis more

12 discernibly than entrectinib in KM12SM-ER cells. These results indicate that foretinib

13 and nintedanib can inhibit the phosphorylation of TRK-A even in the presence of the

14 G667C mutation and suppress the viability of KM12SM-ER cells in vitro.

15 Foretinib overcomes resistance associated with NTRK1-G667C mutation, not

16 NTRK1-G595R mutation.

17 To further assess whether foretinib directly inhibited TRK-A with or without G667C

18 mutation, we examined the effects of foretinib compared with those of entrectinib using

19 Ba/F3, Ba/F3_WT, Ba/F3_G595R, and Ba/F3_G667C cells. Both foretinib and

20 entrectinib inhibited the viability of Ba/F3_WT cells, but not of Ba/F3_G595R cells.

21 Importantly, foretinib, but not entrectinib, inhibited Ba/F3_G667C cell viability (Fig

22 4A). Western blotting revealed that foretinib and entrectinib inhibited the

23 phosphorylation of TRK-A, AKT, and ERK in Ba/F3_WT cells, while entrectinib was

24 more potent than foretinib. However, foretinib, but not entrectinib, inhibited the

1 phosphorylation of TRK-A and AKT in Ba/F3_G667C cells (Fig 4B). These results

2 clearly indicated that foretinib inhibited TRK-A with G667C mutation, while it

3 appeared to be less potent than entrectinib against wild-type TRK-A.

4 Docking analysis of TRK-A with or without G667C mutation and

5 foretinib/entrectinib.

6 We examined whether foretinib could bind to TRK-A with G667C mutation by docking

7 simulation analysis. Inhibitor-binding structures of protein kinase, including TRK-A,

8 are generally divided into DFG-in and DFG-out conformations, which are determined

9 by a flip-flop motion of the highly conserved DFG motif (Asp668-Phe669-Gly670 in

10 TRK-A) (26). The DFG-in conformer is rigid as it is ready for substrate ATP binding,

11 while the DFG-out conformers as auto-inhibition states are divergent, particularly those

12 of TRK-A markedly increase the flexibility of the DFG motif through Gly667.

13 Docking-simulated models based on the crystal structures of the foretinib/EphA2 and

14 entrectinib/Alk complexes (27, 28) support the inhibitory effects of foretinib and

15 entrectinib on the wild-type and G667C mutant of TRK-A. Entrectinib binds to the

16 DFG-in conformation of the wild-type TRK-A and the difluorobenzene group of the

17 inhibitor closely matches and forms a van der Waals contact with the Cα atom of

18 Gly667 (Fig. 4C, left). However, the side chain of the replacing cysteine may be a

19 serious impediment in entrectinib binding to the TRK-A G667C mutant (Fig. 4C, right).

20 It is highly unlikely that the rigid DFG-in conformation is adapted for entrectinib

21 binding to the G667C mutant. Foretinib binds to the DFG-out conformation of the wild

22 type and G667C of TRK-A accompanying the structural adjustment but fits better to

23 G667C compared with the wild-type. The fluorobenzene group of foretinib forms the

24 weakly repulsive interactions with Phe521 of the wild-type but not with that of the

1 G667C mutant (Fig. 4D). Entrectinib and foretinib strongly interact with the C atom of

2 Gly595 in wild type TRK-A. Therefore, the G595R mutation would likely provoke a

3 severe steric hindrance in the binding of both inhibitors to TRK-A (Supplementary Fig.

4 6). This is consistent with the fact that entrectinib and foretinib did not show any

5 inhibitory activity against the G595R mutant. The calculated configuration in the DFG

6 region of TRK-A possesses certain unreliability due to the remarkably higher flexibility.

7 The viability of KM12SM-ER cells was not affected by knockdown of VEGFR-2 or

8 MET.

9 Foretinib and nintedanib are known kinase inhibitors for VEGFR-2/MET and VEGFR-2,

10 respectively (29, 30). To rule out the possibility that the effect of foretinib and

11 nintedanib was mediated by the inhibition of VEGFR-2 or MET, we assessed the effect

12 of knockdown of VEGFR-2 or MET in KM12SM-ER cells. VEGFR-2 protein was

13 undetectable in KM12SM-ER cells (Supplementary Fig. 7B). Treatment with siRNA

14 specific for NTRK1 and MET successfully knocked down TRK-A and MET,

15 respectively (Supplementary Fig. 7D). Under the same experimental conditions,

16 treatment with specific siRNA for VEGFR-2 or MET did not inhibit the viability of

17 KM12SM-ER cells, while the treatment with siRNA for NTRK1 did (Supplementary Fig.

18 7A and 7C). These results strongly suggest that the inhibitory effect of foretinib and

19 nintedanib in KM12SM-ER cells was predominantly due to inhibition of TRK-A, not

20 VEGFR-2 or MET.

21 Foretinib caused the regression of liver metastases produced by KM12SM-ER

22 cells.

23 We next sought to evaluate the effect of such drugs in vivo. Specifically, we decided to

24 focus on foretinib, because, compared with nintedanib, it reduced the viability of

1 KM12SM-ER cells more potently at concentrations close to Cmax (foretinib 90.5 ng/mL

2 [140 nmol/L], nintedanib 34.8 ng/mL [53.6 nmol/L]), as shown in clinical trials (31, 32).

3 The liver is the organ that is affected the most by colon cancer metastases. Therefore,

4 for our in vivo experiments, we first investigated the effect of foretinib in a liver

5 metastasis model obtained through injection of KM12SM-ER cells. After the

6 intra-splenic inoculation of KM12SM-ER cells, bioluminescence in the liver became

7 detectable by day seven. Then, we commenced treatment with entrectinib (15 mg/kg) or

8 foretinib (30 mg/kg), daily, from day 13 to day 22. Entrectinib treatment slightly

9 delayed the progression of liver metastasis but did not decrease the bioluminescence,

10 indicating resistance to entrectinib due to the KM12SM-ER cells in the liver metastasis

11 model (Fig. 5A and B). Under the same experimental conditions, foretinib treatment

12 decreased bioluminescence, indicating the capability of foretinib to reduce liver

13 metastases produced by entrectinib-resistant KM12SM-ER cells (Fig. 5A and B). None

14 of the mice showed significant weight loss (Supplementary Fig. 8A). Western blot

15 analyses of liver metastatic lesions revealed that foretinib, but not entrectinib, inhibited

16 the phosphorylation of TRK-A and AKT in liver tumors, (Fig. 5C), consistent with the

17 data from bioluminescence analysis.

18 We next evaluated tumor cell proliferation and apoptosis in liver metastasis using

19 IHC. The treatment with foretinib remarkably reduced the number of proliferating

20 tumor cells and increased the number of TUNEL-positive tumor cells in liver tumors,

21 compared to the control and entrectinib-treated groups, although the treatment with

22 entrectinib slightly decreased the number of proliferating tumor cells and increased the

23 number of TUNEL-positive tumor cells compared with the control (Fig. 5D and E).

24 These results indicate that foretinib inhibited liver metastases by suppressing tumor cell

1 proliferation and inducing apoptosis of entrectinib-refractory KM12SM-ER cells, and

2 inhibited TRK-A phosphorylation.

3 Foretinib inhibited the progression of brain tumor metastases caused by

4 KM12SM-ER cells.

5 We next evaluated the efficacy of foretinib on brain tumors in a brain

6 metastasis-mimicking model obtained through injection of KM12SM-ER cells. After

7 the intra-cerebral inoculation of KM12SM-ER cells, bioluminescence in the brain

8 became detectable by day seven. When the total flux in the brain lesions reached 1 ×

9 105 photons/sec, we randomized the mice into three groups, and commenced daily

10 treatment with entrectinib (15 mg/kg) or foretinib (30 mg/kg) in two of the groups

11 (treated groups). In both the control and entrectinib groups, bioluminescence increased

12 rapidly, and there was no statistically significant difference between these two groups.

13 Moreover, all mice lost body weight drastically and became moribund by day 14 (seven

14 days after the beginning of the treatment). Under the same experimental conditions,

15 treatment with foretinib remarkably delayed the elevation of bioluminescence,

16 compared with the other two groups (Fig. 6A and B). The body weight loss in the

17 foretinib-treated group was only marginal until day 17 (Supplementary Fig. 8B). These

18 results clearly indicate that foretinib is also effective in the brain metastasis-mimicking

19 model obtained using entrectinib-resistant KM12SM-ER cells. Western blot analyses

20 further revealed that foretinib, but not entrectinib, inhibited the phosphorylation of

21 TRK-A in the brain tumors, consistent with its anti-brain tumor effect observed in this

22 model (Fig. 6C).

23 We next performed IHC analyses on the brain tumors. The treatment with foretinib

24 remarkably reduced the number of proliferating tumor cells and increased the number of

1 TUNEL-positive tumor cells in brain tumors as compared to the untreated group and the

2 entrectinib-treated group. The effect of the entrectinib treatment was only marginal

3 compared with the control (Fig. 6D and E). These results indicate that foretinib

4 inhibited tumor progression by suppressing tumor cell proliferation and inducing

5 apoptosis of entrectinib-refractory KM12SM-ER cells, and inhibiting TRK-A

6 phosphorylation, in the brain metastasis-mimicking model as well as in the liver

7 metastasis model.

9 Discussion

10 Recent reports have described the development of a new compound, LOXO195, which

11 could overcome larotrectinib (a TRK inhibitor)-resistance associated with

12 TRK-A-G595R and TRK-C-G623R in two patients (33). However, the effect of

13 LOXO195 against TRK-A-G667C is weaker than that against TRK-A-G595R. In

14 addition, the effect of LOXO195 on CNS lesions is still unknown. The present study

15 demonstrated three major findings: 1) KM12SM-ER cells acquire entrectinib resistance

16 through the NTRK1-G667C mutation in a brain metastasis-mimicking model; 2) the

17 multiple kinase inhibitor foretinib, which is under evaluation in some clinical trials,

18 inhibits the viability of KM12SM-ER cells bearing the NTRK1-G667C mutation, and is

19 more effective in these cells than in KM12SM cells; 3) foretinib overcomes entrectinib

20 resistance in both liver metastasis and brain metastasis-mimicking models obtained

21 using KM12SM-ER cells. These results suggest that, in tumors expressing NTRK1

22 fusion products, foretinib may be effective in overcoming entrectinib resistance

23 associated with the G667C mutation in tumors of various organs, including the brain.

24 A secondary mutation in a target gene is the most common mechanism of targeted

1 drug resistance (34). Interestingly, in NTRK1, different secondary mutations, such as the

2 G667C and G595R mutations, could be acquired by continuous exposure to entrectinib

3 at different concentrations (11). Specifically, the G667C and G595R mutations were

4 acquired by exposure to low (30–100 nmol/L) and high (1–2 mol/L) concentrations of

5 entrectinib in a KM12 cell model, respectively (11). We speculate that, despite

6 entrectinib is initially effective on patients with CNS lesions, the KM12SM-ER cells

7 with the NTRK1-G667C mutation acquired entrectinib resistance in the brain

8 presumably through the exposure to low concentration of entrectinib because of its

9 relatively lower penetration in the brain compared with that in extra-cranial tumors.

10 This hypothesis is supported by evidence, from the present study, that entrectinib had a

11 marginal effect in a liver metastasis model obtained through injection of KM12SM-ER

12 cells, while the same treatment had no effect in a brain metastasis-mimicking model

13 obtained using the same cell line (Fig. 5B).

14 Several compounds from a kinase inhibitor screening library, including foretinib and

15 nintedanib, which have anti-VEGFR-2 activity, suppressed the viability of

16 KM12SM-ER cells in vitro. Since the KM12SM-ER cells did not express detectable

17 VEGFR-2 protein and treatment with siRNA specific for VEGFR-2 did not affect the

18 viability of KM12SM-ER cells, the observed in vitro effect in KM12SM-ER cells is

19 likely due to the inhibition of a target other than VEGFR-2. Even in the liver metastasis

20 and brain metastasis-mimicking models, the anti-angiogenic effect of foretinib was

21 marginal (Supplementary Fig. 9). On the other hand, foretinib clearly inhibited the

22 phosphorylation of TRK-A in KM12SM-ER cells in vitro and in vivo. Therefore, the

23 antitumor effect of foretinib in our in vivo models obtained through the injection of

24 KM12SM-ER cells may be predominantly due to the inhibition of TRK-A in

1 KM12SM-ER cells. The G667 residue is in the ATP pocket in TRK-A, and the G667C

2 mutation creates steric hindrance that reduces the binding affinity of entrectinib to the

3 TRK-A catalytic pocket (11). It is of interest that foretinib inhibited the viability of

4 KM12SM-ER cells more than KM12SM cells, which strongly suggest that foretinib has

5 higher affinity for the G667C mutant of TRK-A than for the wild type protein. The

6 results of docking simulation supported this hypothesis. Detailed molecular dynamics

7 treatment and/or crystal structure analysis of the foretinib/TRK-A complex would

8 promote further exploration of selective inhibitors for the G667C mutant.

9 Foretinib is under evaluation in clinical trials against various types of solid tumors. In

10 a preclinical study, foretinib was reported to be able to penetrate the BBB (the

11 penetration of the drug in the brain was 14%), and reduce the growth of

12 medulloblastoma cells in an orthotopic CNS dissemination model (35). In line with

13 these results, we demonstrated that foretinib inhibits the tumor progression due to

14 KM12SM-ER cells in a brain metastasis-mimicking model. These findings suggest that

15 foretinib may be effective for the treatment of patients with brain metastases that are

16 entrectinib-resistant and carry the NTRK1-G667C mutation.

17 Recent studies have investigated the sensitivity of different fusion partners to target

18 drugs, using BaF3 or NIH-3T3 cells with target gene transfection. In the case of the

19 RET fusion product, cells expressing KIF5B-RET were shown to have similar sensitivity to the

20 RET-TKIs vandetanib and lenvatinib, when compared to those expressing CCDC6-RET or

21 NCOA4-RET (36, 37). However, a phase II clinical trial investigating the effect of

22 vandetanib in NSCLC with RET fusion genes demonstrated that the effect of vandetanib

23 was different in patients expressing KIF5B-RET or CCDC6-RET, with response rates of

24 20% and 83%, respectively (38), suggesting that the effect of targeted drugs in patients

1 may differ depending on the specific fusion partner in the oncogenic gene

2 rearrangement. The discrepancy of results between the in vitro study and the clinical

3 trial may indicate a limitation of the in vitro model using cells transfected with target

4 genes. In the present study, we performed experiments using two colon cancer cell lines

5 derived from KM12C cells, KM12SM and KM12SM-ER, that express the

6 TMP3-NTRK1 fusion protein, because no other tumor cell line expressing an NTRK1

7 fusion product was available, presumably owing to the low clinical incidence of NTRK1

8 fusion product-positive tumors in cancer patients. The establishment of other tumor cell

9 lines expressing NTRK1 fusion products, possibly with fusion partners other than TMP3,

10 will further clarify the sensitivity and resistance mechanisms of target drugs in tumors

11 expressing NTRK1 fusion products.

12 As indicated in the Introduction, at least four NTRK1 fusion partners, LMNA, TPM3,

13 SQSTM1, and BCAN, have been reported. However, the distinct sensitivity of each

14 NTRK1 fusion product to TRK-A inhibitors is unknown. Additionally, it is unknown

15 whether tumors with different NTRK1 fusion partners acquire resistance to TRK

16 inhibitors through different mechanisms. Russo et al. recently that both

17 LMNA-NTRK1-positive colorectal tumors from patients and the TPM3-NTRK1-positive

18 colon cancer cell line KM12 (same as the KM12C cell line) acquired entrectinib

19 resistance through secondary NTRK1 mutations (G667C and G595R, respectively) (11).

20 In the present study, we also found that TPM3-NTRK1-positive KM12SM cells

21 developed entrectinib resistance that was associated with the G667C mutation in brain

22 tumors. Collectively, these findings suggest that tumor cells with different NTRK1

23 fusion partners acquire TRK inhibitor resistance, at least in part, by common mutations

24 that induce drug resistance. Previous studies have showed that entrectinib is clinically

1 effective for tumors expressing NTRK1 fusion products, including NSCLC expressing

2 SQSTN-NTRK1, glioneural tumors expressing BCAN-NTRK1, and colon cancer

3 expressing LMNA-NTRK1 (5, 6, 39). However, almost all responders suffered from

4 recurrent diseases due to drug resistance (5, 6, 39). Since glioneural tumors produce

5 cranial lesions and NSCLC frequently produces brain metastases, the control of CNS

6 lesions produced by tumors expressing NTRK1 fusion products may become more

7 important in the future.

8 In summary, we demonstrated that entrectinib resistance in brain tumors expressing

9 an NTRK1 fusion product could occur through the NTRK1-G667C mutation. Foretinib

10 can overcome entrectinib resistance associated with the NTRK1-G667C mutation in

11 brain tumors. Our results provide a rationale for clinical trials with foretinib in cancer

12 patients with brain metastases that are entrectinib-resistant and express an NTRK1

13 fusion protein.

16 Acknowledgments

17 The authors thank Dr. I. J. Fidler (M.D. Anderson Cancer Center, Houston, TX) for

18 providing the KM12-SM cells. We also thank Dr. R Katayama (Division of

19 Experimental Chemotherapy, Cancer Chemotherapy Center, Japanese Foundation for

20 Cancer Research, Tokyo, Japan) for providing Ba/F3, Ba/F3_WT, Ba/F3_G595R, and

21 Ba/F3_G667C. We thank Ms. Yumi Akiyama and Ms. Shoko Ueda (Department of

22 Bioinformatics and Genomics, Graduate School of Advanced Preventive Medical

23 Sciences, Kanazawa University) for technical assistance. We thank the Advanced

24 Preventive Medical Sciences Research Center, Kanazawa University for the use of

1 facilities. This study is supported by grants from the Japan Agency for Medical

2 Research and Development, AMED (Grant Numbers 16Ack0106147h0002,

3 17cm0106513h0002 and 17ck0106147h0003 to S. Yano), and JSPS KAKENHI (Grant

4 Number JP16H05308 to S. Yano).

6 References

7 1. Martin-Zanca D, Hughes SH, Barbacid M. A human oncogene formed by the fusion

8 of truncated tropomyosin and protein tyrosine kinase sequences. Nature 1986; 319:

9 743-8.

10 2. Creancier L, Vandenberghe I, Gomes B, Dejean C, Blanchet JC, Meilleroux J, et al.

11 Chromosomal rearrangements involving the NTRK1 gene in colorectal carcinoma.

12 Cancer Lett 2015; 365: 107-11.

13 3. Wiesner T, He J, Yelensky R, Esteve-Puig R, Botton T, Yeh I, et al. Kinase fusions are

14 frequent in Spitz tumours and spitzoid melanomas. Nat Commun 2014; 5: 3116.

15 4. Beimfohr C, Klugbauer S, Demidchik EP, Lengfelder E, Rabes, HM. NTRK1

16 re-arrangement in papillary thyroid carcinomas of children after the Chernobyl

17 reactor accident. Int J Cancer 1999; 80: 842-7.

18 5. Farago AF, Le LP, Zheng Z, Muzikansky A, Drilon A, Patel M, et al. Durable clinical

19 response to entrectinib in NTRK1-rearranged non-small cell lung cancer. J Thorac

20 Oncol 2015; 10: 1670-4.

21 6. Kim J, Lee Y, Cho HJ, Lee YE, An J, Cho GH, et al. NTRK1 fusion in glioblastoma

22 multiforme. PLoS One 2014; 9: e91940.

23 7. Vaishnavi A, Capelletti M, Le AT, Kako S, Butaney M, Ercan D, et al. Oncogenic and

24 drug-sensitive NTRK1 rearrangements in lung cancer. Nat Med 2013; 19: 1469-72.

1 8. Wong V, Pavlick D, Brennan T, Yelensky R, Crawford J, Ross JS, et al. Evaluation of

2 a congenital infantile fibrosarcoma by comprehensive genomic profiling reveals an

3 LMNA-NTRK1 gene fusion responsive to crizotinib. J Natl Cancer Inst 2015; 108:

4 djv307.

5 9. Ardini E, Menichincheri M, Banfi P, Bosotti R, De Ponti C, Pulci R, et al. Entrectinib,

6 a pan-TRK, ROS1, and ALK inhibitor with activity in multiple molecularly deﬁned

7 cancer indications. Mol Cancer Ther 2016; 15: 628-39.

8 10. Drilon A, Siena S, Ou SI, Patel M, Ahn MJ, Lee J, et al. Safety and antitumor

9 activity of the multitargeted pan-TRK, ROS1, and ALK Inhibitor entrectinib:

10 combined results from two phase I trials (ALKA-372-001 and STARTRK-1). Cancer

11 Discov 2017; 7: 400-9.

12 11. Russo M, Misale S, Wei G, Siravegna G, Crisafulli G, Lazzari L, et al. Acquired

13 resistance to the TRK inhibitor entrectinib in colorectal cancer. Cancer Discov 2016;

14 6: 36-44.

15 12. Omuro AM, Kris MG, Miller VA, Franceschi E, Shah N, Milton DT, et al. High

16 incidence of disease recurrence in the brain and leptomeninges in patients with

17 nonsmall cell lung carcinoma after response to gefitinib. Cancer 2005; 103: 2344-8

18 13. Chun SG, Choe KS, Iyengar P, Yordy JS, Timmerman RD. Isolated central nervous

19 system progression on Crizotinib: an Achilles heel of non-small cell lung cancer with

20 EML4-ALK translocation? Cancer Biol Ther 2012; 13: 1376-83.

21 14. Nanjo S, Arai S, Wang W, Takeuchi S, Yamada T, Hata A, et al. MET copy number

22 gain is associated with gefitinib resistance in leptomeningeal carcinomatosis of

23 EGFR-mutant lung cancer. Mol Cancer Ther 2017; 16: 506-15.

24 15. Zhang J, Yu J, Sun X, Meng X. Epidermal growth factor receptor tyrosine kinase

1 inhibitors in the treatment of central nerve system metastases from non-small cell

2 lung cancer. Cancer Lett 2014; 351: 6-12.

3 16. Yano S, Shinohara H, Herbst RS, Kuniyasu H, Bucana CD, Ellis LM, et al.

4 Expression of vascular endothelial growth factor is necessary but not sufficient for

5 production and growth of brain metastasis. Cancer Res 2000; 60: 4959-67.

6 17. Zheng Z, Liebers M, Zhelyazkova B, Cao Y, Panditi D, Lynch KD, et al. Anchored

7 multiplex PCR for targeted next-generation sequencing. Nat Med 2014; 20: 1479-84.

8 18. Tevlin R, Larkin JO, Hyland JM, O’Connell PR, Winter DC. Brain metastasis from

9 colorectal carcinoma: a single cancer centre experience. Ir J Med Sci 2015; 184:

10 673-5.

11 19. Morikawa K, Walker SM, Nakajima M, Pathak S, Jessup JM, Fidler IJ. Influence of

12 organ environment on the growth, selection, and metastasis of human colon

13 carcinoma cells in nude mice. Cancer Res 1988; 48: 6863-71.

14 20. Kim D, Pertea G, Trapnell C, Pimental H, Kelley R, Salzberg SL. TopHat2: accurate

15 alignment of transcriptomes in the presence of insertions, deletions and gene fusions.

16 Genome Biol 2013; 14: R36.

17 21. Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, et al.

18 Integrative genomics viewer. Nat Biotechnol 2011; 29: 24-6.

19 22. Daniel N, Mihaela S, Henrik E, Sara K, Astrid M, Olli K, et al. FusionCatcher - a

20 tool for finding somatic fusion genes in paired-end RNA-sequencing data. bioRxiv

21 2014; doi: https://doi.org/10.1101/011650.

22 23. Nanjo S, Nakagawa T, Takeuchi S, Kita K, Fukuda K, Nakada M, et al. In vivo

23 imaging models of bone and brain metastases and pleural carcinomatosis with a novel

24 human EML4-ALK lung cancer cell line. Cancer Sci 2015; 106: 244-52.

1 24. Moreno A, Lopez LA, Fabra A, Arus C. 1H MRS markers of tumor growth in

2 intrasplenic tumors and liver metastasis induced by injection of HT-29 cells in nude

3 mice spleen. NMR Biomed 1998; 11: 93-106.

4 25. Takeuchi S, Fukuda K, Arai S, Nanjo S, Kita K, Yamada T, et al. Organ-specific

5 efficacy of HSP90 inhibitor in multiple-organ metastasis model of chemorefractory

6 small cell lung cancer. Int J Cancer 2016; 138: 1281-9.

7 26. Wu P, Nielsen TE, Clausen MH. FDA-approved small-molecule kinase inhibitors.

8 Trends Pharmacol Sci 2015; 36: 422–39.

9 27. Heinzlmeir S, Kudlinzki D, Sreeramulu S, Klaeger S, Gande SL, Linhard V, et al.

10 Chemical proteomics and structural biology define EPHA2 inhibition by clinical

11 kinase drugs. ACS Chem Biol 2016; 11: 3400–11.

12 28. Menichincheri M, Ardini E, Magnaghi P, Avanzi N, Banfi P, Bossi R, et al.

13 Discovery of entrectinib: A new 3-aminoindazole as a potent anaplastic lymphoma

14 kinase (ALK), c-ros oncogene 1 kinase (ROS1), and pan-tropomyosin receptor

15 kinases (Pan-TRKs) inhibitor. J Med Chem 2016; 59: 3392–408.

16 29. Shah MA, Wainber ZA, Catenacci DV, Hochster HS, Ford J, Kunz P, et al. Phase II

17 study evaluating 2 dosing schedules of oral foretinib (GSK1363089),

18 cMET/VEGFR2 inhibitor, in patients with metastatic gastric cancer. PLoS One 2013;

19 8: e54014.

20 30. Rolfo C, Raez LE, Bronte G, Santos ES, Papadimitriou K, Buffoni L, et al. BIBF

21 1120/nintedanib: a new triple angiokinase inhibitor-directed therapy in patients with

22 non-small cell lung cancer. Expert Opin Investig Drugs 2013; 22: 1081-8.

23 31. Eder JP, Shapiro GI, Appleman LJ, Zhu AX, Miles D, Keer H, et al. A phase 1 study

24 of foretinib, a multi-targeted inhibitor of c-Met and vascular endothelial growth factor

1 receptor 2. Clin Cancer Res 2010; 16: 3507-16.

2 32. Mross K, Stefanic M, Gmehling D, Frost A, Baas F, Unger C, et al. Phase 1 study of

3 the angiogenesis inhibitor BIBF 1120 in patients with advanced solid tumors. Clin

4 Cancer Res 2010; 16: 311-9.

5 33. Drilon A, Nagasubramanian R, Blake JF, Ku N, Tuch BB, Ebata K, et al. A

6 next-generation TRK kinase inhibitor overcomes acquired resistance to prior TRK

7 kinase inhibition in patients with TRK fusion-positive solid tumors. Cancer Discov

8 2017; 7: 963-72.

9 34. Balak MN, Gong Y, Riely GJ, Somwar R, Li AR, Zakowski MF, et al. Novel

10 D761Y and common secondary T790M mutations in epidermal growth factor

11 receptor-mutant lung adenocarcinomas with acquired resistance to kinase inhibitors.

12 Clin Cancer Res 2006; 12: 6494-501.

13 35. Faria CC, Golbourn BJ, Dubuc AM, Remke M, Diaz RJ, Aqnihotri S, et al. Foretinib

14 is effective therapy for metastatic sonic hedgehog medulloblastoma. Cancer Res

15 2015; 75: 134-46.

16 36. Kodama T, Tsukaguchi T, Satoh Y, Yoshida M, Watanabe Y, Kondoh O, et al.

17 Alectinib shows potent antitumor activity against RET-rearranged non-small cell lung

18 cancer. Mol Cancer Ther 2014; 13: 2910-8.

19 37. Okamoto K, Kodama K, Takase K, Sugi NH, Yamamoto Y, Iwata M, et al.

20 Antitumor activities of the targeted multi-tyrosine kinase inhibitor lenvatinib (E7080)

21 against RET gene fusion-driven tumor models. Cancer Lett 2013; 340: 97-103.

22 38. Yoh K, Seto T, Satouchi M, Nishio M, Yamamoto N, Murakami H, et al. Vandetanib

23 in patients with previously treated RET-rearranged advanced non-small-cell lung

24 cancer (LURET): an open-label, multicentre phase 2 trial. Lancet Respir Med 2017;

1 5: 42-50.

2 39. Sartore-Bianchi A, Ardini E, Bosotti R, Amatu A, Valtorta E, Somaschini A, et al.

3 Sensitivity to entrectinib associated with a novel LMNA-NTRK1 gene fusion in

4 metastatic colorectal cancer. J Natl Cancer Inst 2015; 108: djv306.

1 Figure Legends

2 Figure 1. Establishment of KM12SM-ER cells with acquired resistance to

3 entrectinib in vivo.

4 (A) We transfected the EGFP/Eluc gene into KM12SM cells to establish KM12SM/Eluc

5 cells. KM12SM/Eluc cells were inoculated into the brain of SCID mice. The mice were

6 treated daily with or without entrectinib (15 mg/kg) for 37 days until the

7 bioluminescence increased. Mean ± SE of total flux are shown in the lower panel. Then,

8 the entrectinib-treated brain tumor was harvested at the point indicated by the orange

9 triangle and cultured in vitro. The expanded tumor cells were named KM12SM-ER. (B)

10 The sensitivity of KM12SM-ER and KM12SM cells to entrectinib was determined

11 through cell viability assays, using a CCK-8 kit. The data (mean ± standard deviation

12 [SD] of triplicate cultures) shown are representative of three independent experiments

13 with similar results. (C) Tumor cells were treated with entrectinib (10 nmol/L) for 4 h or

14 c-PARP for 48 h, and harvested lysates were assessed by western blotting. Data shown

15 are representative of three independent experiments with similar results. (D) KM12SM

16 and KM12SM-ER cells were treated with siRNA specific for NTRK1 (siNTRK) or a

17 scrambled control (siSCR). Then, cell viability was determined with a CCK-8 kit. Data

18 (mean ± SD) shown are representative of three independent experiments, with similar

19 results. *P < 0.05, as determined by the Student's t-test, compared with siSCR. (E)

20 KM12SM and KM12SM-ER cells were transfected with NTRK1-specific siRNA

21 (siNTRK) or a scrambled control (siSCR). After 48 h, the cell lysates were harvested

22 and evaluated for protein expression by western blotting. Three independent

23 experiments were performed.

1 Figure 2. KM12SM-ER cells harbored the NTRK1-G667C mutation.

2 (A) Schematic representation of RNA-sequencing reads indicating the presence of

3 TPM3-NTRK1 fusion transcripts. FusionCatcher software was used to count the number

4 of paired-end reads relative to fusion transcripts (‘Spanning pairs’) or mapping on the

5 fusion junction (‘Spanning unique reads’). The red-colored transverse line indicates the

6 RT-PCR target region, encompassing the TPM3-NTRK1 fusion junction and the site of

7 the acquired p.G667C mutation (denoted by the asterisk). (B) RT-PCR followed by

8 agarose gel electrophoresis confirmed the presence of TPM3-NTRK1 fusion transcripts

9 in KM12SM/Eluc and KM12SM-ER cells. (C) Sanger sequencing of the RT-PCR

10 products showed the identity of the fusion junctions in the TPM3-NTRK1 fusion

11 transcripts from the two cell lines. (D) Integrative Genomics Viewer images illustrating

12 mapped RNA sequencing reads, which indicate the presence of the NTRK1 sequence

13 alteration p.G667C (NM_002529.3: c.1999G>T) in KM12SM-ER cells (lower panel),

14 but not KM12SM/Eluc cells (upper panel). (E) Sanger sequencing electropherograms of

15 RT-PCR products validating the acquired mutation p.G667C in the TPM3-NTRK1

16 fusion transcripts only in KM12SM-ER cells.

18 Figure 3. Screening of compounds that inhibit the viability of KM12SM-ER cells.

19 (A) The effect of 122 compounds from a Kinase Inhibitor Library on the viability of

20 KM12SM-ER cells was assessed using a CCK-8 kit. The 122 kinase inhibitors (100

21 nmol/L) are rank-ordered from the highest to the lowest according to their inhibitory

22 effect on the viability of KM12SM-ER cells. The effects of the top 13 compounds

23 indicated by a red line with an arrow head are further shown in (B). (B) Red bars

24 indicate the compounds that are known to have anti-VEGFR-2 activity. (C) The effect

1 of entrectinib, nintedanib, and foretinib on the viability of KM12SM and KM12SM-ER

2 cells was determined with a CCK-8 kit. Data (mean ± SD of triplicate cultures) shown

3 are representative of three independent experiments with similar results. (D) KM12SM

4 and KM12SM-ER cells were incubated with the indicated drugs (10 nmol/L) for 4 h or

5 48 h (for c-PARP), and the cell lysates were harvested. The expression of indicated

6 molecules was determined by western blot analysis. Data shown are representative of

7 three independent experiments with similar results.

9 Figure 4. Sensitivity of Ba/F3, Ba/F3_WT, Ba/F3_G595R, and Ba/F3_G667C cells

10 to entrectinib and foretinib and structural analysis of TRK-A with or without

11 G667C mutation and foretinib/entrectinib.

12 (A) The sensitivity of Ba/F3, Ba/F3_WT, Ba/F3_G595R, and Ba/F3_G667C cells to

13 foretinib or entrectinib was determined through cell viability assays using a CCK-8 kit.

14 The data [mean ± standard deviation (SD) of triplicate cultures] shown are

15 representative of three independent experiments with similar results. (B) Ba/F3_WT and

16 Ba/F3_G667C cells were incubated with the indicated drugs (3 nmol/L) for 4 h and the

17 cell lysates were harvested. The expression of indicated molecules was determined by

18 western blot analysis. Data shown are representative of three independent experiments

19 with similar results. (C) Binding structures of entrectinib with the wild-type (left) and

20 TRK-A with G667C mutation (right). (D) Binding structures of foretinib with the

21 wild-type (left) and TRK-A with G667C mutation (right).

22 Figure 5. Therapeutic effect of foretinib and entrectinib in a liver metastasis model

23 obtained using KM12SM-ER cells.

24 (A) KM12SM-ER cells were inoculated into the spleen of SCID mice, and the mice

1 were randomized on day 13. Mice were treated orally with entrectinib (15 mg/kg) or

2 foretinib (30 mg/kg) daily, from day 13 to 21. Bioluminescence was evaluated twice a

3 week, from day 7 to day 22. Data shown are mean ± standard error (SE) for five mice in

4 each group. (B) Representative bioluminescent and photographic merged images of

5 mice. (C) Mice were sacrificed on day 22 and liver metastatic lesions were harvested.

6 The protein lysates extracted from the liver lesions were used for western blot assays to

7 detect the expression of the indicated molecules. (D) Proliferating tumor cells (Ki-67)

8 and apoptotic cells (TUNEL) in the liver metastases were identified by IHC. Bar, 100

9 µm. (E) Quantification of proliferating and apoptotic cells in liver metastases,

10 determined by IHC in D. The data shown are the mean ± SD of five areas. *P < 0.05 as

11 determined with the Mann-Whitney U test.

13 Figure 6. Therapeutic effect of foretinib and entrectinib in a brain

14 metastasis-mimicking model obtained using KM12SM-ER cells.

15 (A) KM12SM-ER cells were inoculated into the brain of SHO mice, and the mice were

16 randomized on day seven. Mice were treated orally with entrectinib (15 mg/kg) or

17 foretinib (30 mg/kg) daily, from day 7 to 24. Bioluminescence was evaluated twice a

18 week, from day 7 to day 24. Data shown are mean ± SE for five mice in each group. (B)

19 Representative bioluminescent and photographic merged images of mice. (C) When

20 mice became moribund, they were sacrificed and brain tumor lesions were harvested.

21 The protein lysates extracted from the brain lesions were used for western blots to detect

22 expression of the indicated molecules. (D) Proliferating tumor cells (Ki-67) and

23 apoptotic cells (TUNEL) in the brain metastases were identified by IHC. Bar, 100 µm.

24 (E) Quantification of proliferating and apoptotic cells in brain metastases, determined

1 by IHC in D. The data shown are the mean ± SD of five areas. *P < 0.05 as determined

2 with the Mann-Whitney U test.

Foretinib overcomes entrectinib resistance associated with the NTRK1 G667C mutation in NTRK1 fusion-positive tumor cells in a brain metastasis model

Akihiro Nishiyama, Tadaaki Yamada, Kenji Kita, et al.

Clin Cancer Res Published OnlineFirst February 20, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-17-1623

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2018/02/20/1078-0432.CCR-17-1623.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet Manuscript been edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2018/02/20/1078-0432.CCR-17-1623. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.