Downloaded from Bioscientifica.Com at 09/23/2021 08:03:31PM Via Free Access

25 5

Endocrine-Related E T Aristizabal Prada, Tropomyosin receptor kinase 25:5 547–560 Cancer V Heinzle et al. (Trk) in NETs RESEARCH Tropomyosin receptor kinase: a novel target in screened neuroendocrine tumors

Elke Tatjana Aristizabal Prada1,2,3,*, Vera Heinzle1,2,*, Thomas Knösel1,4, Svenja Nölting1,2,3, Gerald Spöttl1,2,3, Julian Maurer1,2,3, Christine Spitzweg1,2,3, Martin Angele1,5, Nina Schmidt1,2,3, Felix Beuschlein3,6, Günter K Stalla7, Rainer Blaser8, Klaus A Kuhn8 and Christoph J Auernhammer1,2,3

1Interdisciplinary Center of Neuroendocrine Tumors of the GastroEnteroPancreatic System (GEPNET-KUM), Klinikum der Universität München, Ludwig-Maximilians-University of Munich, Campus Grosshadern, Munich, Germany 2Department of Internal Medicine 2, University-Hospital, Klinikum der Universität München, Ludwig-Maximilians-University of Munich, Munich, Germany 3Department of Internal Medicine 4, University-Hospital, Klinikum der Universität München, Ludwig-Maximilians-University of Munich, Munich, Germany 4Institute of Pathology, Ludwig-Maximilians-University of Munich, Munich, Germany 5Department of General, Visceral, Transplantation, Vascular and Thoracic Surgery, University-Hospital, Klinikum der Universität München, Ludwig- Maximilians-University of Munich, Munich, Germany 6Klinik für Endokrinologie, Diabetologie und Klinische Ernährung, Universitätsspital Zürich, Zurich, Switzerland 7Clinical Neuroendocrinology, Max Planck Institute of Psychiatry, Munich, Germany 8Institute of Medical Statistics and Epidemiology, Klinikum Rechts der Isar, Technische Universität München, Munich, Germany

Correspondence should be addressed to C J Auernhammer: [email protected]

*(E T Aristizabal Prada and V Heinzle contributed equally to this work)

Abstract

Tropomyosin receptor kinase (Trk) inhibitors are investigated as a novel targeted therapy Key Words in various cancers. We investigated the in vitro effects of the pan-Trk inhibitor GNF-5837 ff tropomyosin kinase in human neuroendocrine tumor (NET) cells. The human neuroendocrine pancreatic receptor (Trk) BON1, bronchopulmonary NCI-H727 and ileal GOT1 cell lines were treated with GNF-5837 ff Trk inhibitor alone and in combination with everolimus. Cell viability decreased in a time- and dose- ff neuroendocrine tumor dependent manner in GOT1 cells in response to GNF-5837 treatment, while treatment in ff everolimus BON1 and NCI-H727 cells showed no effect on cellular viability. Trk receptor expression determined GNF-5837 sensitivity. GNF-5837 caused downregulation of PI3K-Akt-mTOR signaling, Ras-Raf-MEK-ERK signaling, the cell cycle and increased apoptotic cell death. The combinational treatment of GNF-5837 with everolimus showed a significant enhancement in inhibition of cell viability vs single substance treatments, due to a cooperative PI3K-Akt-mTOR and Ras-Raf-MEK-ERK pathway downregulation, as well as an enhanced cell cycle component downregulation. Immunohistochemical staining for Trk receptors were performed using a tissue microarray containing 107 tumor samples of gastroenteropancreatic NETs. Immunohistochemical staining with TrkA receptor and pan-Trk receptor antibodies revealed a positive staining in pancreatic NETs in 24.2% (8/33) and 33.3% (11/33), respectively. We demonstrated that the pan-Trk inhibitor GNF-5837 has promising anti-tumoral properties in human NET cell lines expressing the TrkA receptor. Immunohistochemical or molecular screening for Trk expression particularly in pancreatic NETs might serve as predictive marker for molecular targeted Endocrine-Related Cancer therapy with Trk inhibitors. (2018) 25, 547–560

10.1530/ERC-17-0201 Endocrine-Related E T Aristizabal Prada, Tropomyosin receptor kinase 25:5 548 Cancer V Heinzle et al. (Trk) in NETs

Introduction constitutively activated Trk family kinase have been reported in various cancer types (Amatu et al. 2016). Neuroendocrine tumors (NETs) are highly heterogeneous Recently, Trk inhibitors have emerged as promising tumors originating from distinct cell precursors (Schimmack novel targets in molecular targeted cancer therapy (Amatu et al. 2011). Current data show an increase in incidence et al. 2016, Demir et al. 2016, Khotskaya et al. 2017). In a of gastroenteropancreatic neuroendocrine tumors large-cell neuroendocrine carcinoma (LCNEC) cell line the (GEP-NETs) lately (Modlin et al. 2008a, Yao et al. 2008). pan-Trk inhibitor AZD7451 showed promising Despite increasing knowledge on the tumor biology, pharmacologic potential, particularly in cells with a NTRK genetics and epigenetics of GEP-NETs, the overall outcome gene fusion with constitutive Trk expression (Tatematsu et al. and survival have gained very little over the last decades 2014). Furthermore, BDNF/TrkB expression was significantly (Modlin et al. 2008b, Kidd et al. 2016, Karpathakis et al. upregulated in clinical LCNEC specimens and has shown to 2017). Peptide receptor radionuclide therapy has shown be implicated in tumorigenicity and invasiveness in vitro and promising results regarding progression-free and overall in vivo (Odate et al. 2013). TrkB signaling has been associated survival (van Vliet et al. 2013, Bodei et al. 2014), but other with cellular survival and migration in neuroendocrine lung systemic therapeutic approaches for GEP-NETs, such as carcinomas (Osborne et al. 2013). biotherapy, chemotherapy and molecular targeted therapy Several Trk inhibitors, such as LOXO-101 and are limited in their efficiency (Auernhammer & Goke 2011, entrectinib (RXDX-101), are currently investigated in Pavel & Sers 2016, Cives & Strosberg 2017). Currently the clinical phase 1/2 trials and have shown encouraging only approved molecular targeting strategies for NETs are the activity including patients with molecularly defined mTORC1 inhibitor everolimus (Pavel et al. 2011, Yao et al. cancers with rearrangements of Trk family kinases 2011) and the multi-tyrosin kinase inhibitor sunitinib (Ardini et al. 2016, Passiglia et al. 2016, Drilon et al. 2017). (Raymond et al. 2011) demonstrating the urgent need for So far, no data on the putative role of Trk inhibitors innovative molecular therapeutic targets and strategies. in the treatment of GEP-NETs have been available. Here, As recently reviewed by Amatu et al. (2016), the Trk we demonstrate anti-proliferative effects of the highly family TrkA, TrkB and TrkC is encoded by the NTRK genes specific pan-Trk inhibitor GNF-5837 Albaugh( et al. 2012) NRTK1, NTRK2 and NTRK3, respectively. Neurotrophins in the human neuroendocrine tumor cell line GOT1 (NTs) are specific ligands to the Trk receptors, especially in vitro and examine TrkA downstream signal transduction. nerve growth factor (NGF) for TrkA, brain-derived In addition, we investigated the expression profile of Trk growth factor (BDGF) and NT-4/-5 for TrkB and NT-3 in GEP-NET tumor samples from patients. for TrkC (Kaplan et al. 1991, Klein et al. 1991, Barbacid 1994, Bibel & Barde 2000, Huang & Reichardt 2003, Amatu et al. 2016). All Trks are transmembrane receptors Materials and methods and ligand binding to the extracellular domain causes Materials receptor oligomerization and autophosphorylation of the intracellular domain containing the kinase GNF-5837 was purchased from R&D Systems and region. Subsequently, activated downstream effectors recombinant human NGF was purchased from Life of the Trk signaling involve the Ras-Raf-MEK-ERK, the Technologies. Human BDNF was obtained from Sigma- PI3K-Akt-mTOR, the PLCγ-1 and the PKC pathway (Huang Aldrich and everolimus (07741FLUKA) was purchased & Reichardt 2003, Amatu et al. 2016). from Sigma. These substances were diluted in dimethyl- The Trk receptor family has been demonstrated sulfoxide (DMSO, 10 mM stock solution; Sigma, D8418). to be implicated in the growth of many different types Dulbecco’s Modified Eagle medium–Nutrient Mixture of cancer, for example, oral squamous cell carcinoma F-12, 1:1 (DMEM/F12) as well as penicillin/streptomycin (McGregor et al. 1999), pancreas (McGregor et al. 1999), were obtained from Gibco/Invitrogen. RPMI medium thyroid (McGregor et al. 1999), lung (Oelmann et al. (with l-Glutamine, NaCO3) and phosphate-buffered 1995), esophagus (Zhu et al. 2000), prostate (Sortino et al. saline (PBS) were purchased from Sigma, whereas Trypsin- 2000), breast (Tagliabue et al. 2000) and ovarian cancer EDTA (10×) was acquired from PAA Laboratories (Cölbe, (Davidson et al. 2001) and glioblastoma (Oelmann et al. Deutschland). Fetal bovine serum (FBS) and amphotericin 1995). Chromosomal rearrangements with oncogenic B were acquired from Biochrom (Berlin, Germany). gene fusions joining the 3′ region of the NTRK1, NTRK2 Human adult normal brain tissue Lysate (ab29466) was or NTRK3 gene to another partner gene resulting in a obtained from Abcam and stored at −80°C.

Cell culture medium. Then, the medium was replaced by fresh 10% FBS medium and cells were incubated with The human pancreatic neuroendocrine BON1 tumor cell different concentrations of GNF-5837 (GNF), NGF line (Evers et al. 1991) (kindly provided by Prof. R Göke, and everolimus (RAD001). The incubation times were Marburg, Germany) was grown in DMEM/F12 (1:1) from 10 min up to 72 h. Short times of incubation supplemented with 10% FBS, 1% penicillin/streptomycin (10 min) were needed to see the effects on Trk receptor and 0.4% amphotericin B and passages (p) from p7 up stimulation (NGF), whereas longer times of incubation to p20 were used for the main experiments. Human (>2 h) were needed to see the effects on Trk receptor bronchopulmonary neuroendocrine NCI-H727 tumor cells inhibition (GNF-5837). After incubation, the cells were (p11 – p23 for main experiments) (Cakir & Grossman 2011) washed twice in cold PBS on ice and lysed in 200 μL (acquired from ATCC) and human midgut carcinoid GOT1 Lysis buffer (M-PER 1 Mammalian Protein Extraction cells (Kolby et al. 2001) (kindly provided by Prof. O Nilsson, Reagent containing HALT protease & phosphatase Sahlgrenska University Hospital Göteborg, Sweden) were inhibitor cocktail, Thermo Scientific). Lysates were grown in RPMI medium supplemented with 10% FBS, 1% centrifuged at 16,060 g for 10 min. Supernatants penicillin/ streptomycin and 0.4% amphotericin B. The were adjusted to the same protein concentration exact passage status of GOT1 cells is not known, as there is no (30–50 μg/50 μL) (Rotiquant Universal, CarlRoth, adequate information from the original institute available. Karlsruhe, Germany) and denatured in sodium dodecyl GOT1 culture medium was additionally supplemented sulfate (SDS) sample buffer (0.25% Tris–HCL, 40% with 5 g/mL apo-transferrin and 0.135 IU/mL μ glycerol, 2% SDS, 1% dithiothreitol, bromophenol insulin. The doubling time of NET cell lines used in this blue, pH 8.8). Equal amounts of protein were separated study are: 18 day in GOT1, 42 h in NCI-H727 and 20 h in ≥ on a SDS polyacrylamide gel and electrotransferred for BON1 cells (Grozinsky-Glasberg et al. 2012). 60 min onto PVDF membranes (Immobilon; Millipore) All human neuroendocrine cell lines were received using a semi-dry Western blot technique. After blocking and cultured, as recently described (Zitzmann et al. 2007, in 2% skimmed milk powder, the membranes were Spampatti et al. 2014). The cells were tested mycoplasma incubated overnight at 4°C in appropriate dilutions of free and incubated at 37°C and 5% CO . All cell lines were 2 primary antibodies against pTrkA (Tyr 674/5) 1:1000, authenticated by the German Biological Resource Centre pTrkB (Tyr 706/7) (#4621) 1:1000, TrkA (#2508) 1:1000, DSMZ (DSMZ, Braunschweig, Germany) using short Trkpan (#4609) 1:1000, pAKT (Ser473) (#4060) 1:20,000, tandem repeats (STRs) analysis. AKT(#2920) 1:10,000, pERK (Thr202/Tyr204)1/2 (#4370) 1:10,000, p4EBP1 (Ser65) (#9451) 1:2000, 4EBP1 (#9644) Cell viability assessment 1:5000; pGSK3 (Ser21/9) (#9331) 1:1000, GSK3 (#5676) 1:5000, pp70S6K (Thr389) (#9234) 1:1000, p70S6K Cells were counted (Cell Counter Countess, Invitrogen) (#9202) 1:1000, pS6 (Ser240/4) (#5364) 1:25,000, S6 seeded and grown for 24 h in 96-well plates at densities (#2317) 1:2000, pEGFR (Tyr1068) (#3777) 1:1000, EGFR of 1500 (BON1), 2000 (NCI-H727) and 30,000 (GOT1) (#4267) 1:5000, Cyclin D3 (#2936) 1:2000 (all from cells per well. The following day the cells were treated Cell Signaling), Actin (A5441) 1:20,000 (Sigma), ERK1/2 in 10% FBS medium with different concentrations of (06–182) (Merck-Millipore) 1:20,000, Bcl2 (ab694) GNF-5837 alone and in combination with 10 nM everolimus, 1:1000, TrkB (ab134155) 1:1000, TrkC (ab117966) 1:500 as well as with NGF. Metabolic activity was measured by (all from Abcam), p21 Waf1/Cip1 (BD610233) 1:2000 ‘Cell Titer96Aqueous OneSolution’ cell viability assay (BD Biosciences, Heidelberg, Germany), p53 (sc-126) (Promega) after 72-, 96- and 144-h of incubation. After 4 h 1:20000 (Santa Cruz) Anti-NGF (EP1320Y) (ab52918) of incubation with CellTiter 96solution, the absorbance at 1:2000 (Abcam). After washing in TBS, the membranes 490 nm was stated using the ELISA plate reader (Orion II, were incubated with a peroxidase conjugated Berthold Detection Systems, Pforzheim, Germany). The secondary antibody (Anti-Rabbit-IgG, HRP-linked highest DMSO concentration used in our experimental set Antibody, Cell Signaling, #7074 or Anti-Mouse-IgG, up was aprox. 700 µM (0.005%). HRP-linked Antibody, Cell Signaling, #7076) 1:25,000 for 2 h. The blots were washed and immersed in the Protein extraction and Western blotting chemiluminescent substrate Super Signal West Dura For Western blot experiments, 1 × 106 (GOT1) were (Thermo Scientific), and images were taken with an ECL seeded in 6-well plates and grown for 24 h in complete ChemoCam Imager (INTAS, Göttingen, Germany).

Cell cycle analysis by flow cytometric Statistical analysis analysis (FACS) The results are displayed as mean ± standard deviation of Cell cycle phase distribution was analyzed using the mean (s.d.) of at least three independently performed propidium iodide staining and flow cytometry (BD Accuri experiments. Each cell viability experiment consisted C6 Analysis, BD Biosciences). Cells were cultured in 6-well of at least 6 samples per substance concentration and plates for 24 h. Subsequently, medium was replaced by incubation period. Some a priori tests considering the fresh medium and cells were incubated with different normal distribution and homogeneity of variances concentrations of GNF-5837. After 72 h, cells were washed were performed applying the Kolmogorov–Smirnov with PBS and treated with 300 μL trypsin for 5 min at 37°C. test and the Levene’s test of the SPSS statistical package Cells were collected, washed and resuspended in 300 μL SPSS (version 13.0 for Windows, SPSS (2005)). When propidium iodide solution (Sigma-Aldrich). parametric criteria were met an ANOVA comparison of means with a post hoc Tukey test or a two-tailed t-test was performed; when non-parametric criteria were met the Patient cohort and generation of a tissue Kruskal–Wallis followed by the Mann–Whitney test was microarray (TMA) performed. Statistical significance was assessed at P < 0.05. Patients with NET of the gastroenteropancreatic system For the immunohistochemistry a Pearson chi-square test treated at the Klinikum der Universität München were was performed. enrolled in the NeoExNET (Network of Excellence for Neuroendocrine Tumors in Munich) program. The NeoExNET program was approved by the Local Ethics Results Committee of the University of Munich and by the GNF-5837 effectively inhibits GOT1 cell growth data safety manager (Datenschutzbeauftragter) of the University of Munich, and all patients had given written Cell viability assay determined a clear decrease of GOT1 informed consent. Paraffin-embedded tissue samples from cell viability in vitro in response to the pan-TRK inhibitor pathology specimens of patients enrolled in the NeoExNET GNF-5837 in a time- and dose-dependent manner (Fig. 1 program were used to generate a tissue microarray (TMA) and Table 1). At all three tested incubation periods of GEP-NETs. The TMA encompassed tumor samples (48, 96 and 144 h), significant differences when compared from 107 patients. Histology was reviewed in all tumors to untreated control cells (control DMSO) were detected to confirm the diagnosis. The specimens were classified (P < 0.05) (Fig. 1). At a GNF-5837 concentration of 50 nM, to the new WHO 2010 classification scheme and tumor a plateau of dose efficacy was reached, where higher grading (G) was assessed (‘G1’ = 48 patients, ‘G2’ = 51 doses did not yield a better effect in cell survival decrease patients, ‘G3’ = 3 patients and ‘no information’ = 5 (Fig. 1). IC20 and IC50 inhibitory concentrations of patients). A tissue microarray was assembled using 0.6 mm GNF-5837 in GOT1 cells were calculated (Table 1). punch biopsies from all samples according to standard procedures (Knosel et al. 2005, 2006). Anti-proliferative effect of the Trk inhibitor GNF-5837 is associated with functional Trk Immunohistochemical staining of the TMA expression in human neuroendocrine tumor cells

Immunohistochemical staining was performed using TrkA GNF-5837 was also tested on the pulmonary carcinoid cell (#2508) and Trkpan (#4609) 1:100 from Cell Signaling. line NCI-H727 and the pancreatic neuroendocrine tumor Slides were pre-treated by cooking, and then incubated cell line BON1, however, without any effects on cellular with the antibodies, followed by antibody detection using growth inhibition (Supplementary Fig. 1, see section on a biotinylated anti-mouse secondary antibody and a supplementary data given at the end of this article). In multilink biotin-streptavidin-amplified detection system relation with this observed primary GNF-5837 resistance, (Biogenex, San Ramon, CA, USA). A Fastred chromogen Western blot analysis failed to detect any expression of system (DAKO) was used to visualize staining. A 4-tier TrkA, TrkB or TrkC receptor in BON1 and NCI-H727 cell scale was used for a semi-quantitative score (0 = negative, protein extracts (Fig. 2A). In contrast, protein extracts 1 = weak expression, 2 = moderate expression, 3 = strong from human midgut carcinoid GOT1 cells showed expression). significant levels of TrkA expression (145 kDa), faint TrkC

Figure 1 Cell survival of GOT1 cells with different concentrations of the pan-TRK inhibitor GNF-5837 (from 0.1 nM to 500 nM) and different times of incubation (48, 96 and 144 h). The arithmetic means and s.d. of at least three independent experiments are shown. Statistically significant different results in comparison to Control DMSO considering ***P < 0.001. expression (145 kDa) and no TrkB expression (Fig. 2A). A specificity of the ligand NGF to its receptor TrkA (Amatu strong signal was also found in protein extracts derived et al. 2016). Incubation of GOT1 cells with NGF caused from GOT1 cells, using a pan-Trk antibody detecting a rapid phosphorylation of TrkA receptor after 10 min, TrkA, TrkB and TrkC (Fig. 2A). which was partially inhibited by co-incubation with the A human adult normal brain tissue lysate (ab29466 Trk inhibitor GNF-5837 (Fig. 3B). The phosphorylation from Abcam) served as positive control for all three status of TrkA decreased continuously with increasing specific Trk receptor antibodies Numakawa( et al. 2010) concentration of the pan-TRK inhibitor GNF-5837 (5, 50 (Fig. 2A). The different weights of Trk in the human brain and 500 nM) (Fig. 3C). Most prominent dephosphorylation tissue (90 kDa) is explained by either alternative splicing effects were reached at the highest dose of the inhibitor forms of Trk (Clary & Reichardt 1994) or an immature (Fig. 3C). All time periods of incubation showed the form of the receptor (Martin-Zanca et al. 1989) (Fig. 2A). same dose-dependent dephosphorylation pattern of TrkA Next, we investigated the expression levels of the TrkA (Fig. 3C). Interestingly, after long periods of incubation ligand NGF in all three cell lines. While NGF expression (24, 48 and 72 h) with the inhibitor GNF-5837 the was detectable in human brain tissue (positive control), unphosphorylated TrkA receptor shows a compensatory BON1 and NCI-H727 cells, in GOT1 cells only a low upregulation (Fig. 3C). intensity band of NGF could be detected by Western blot analysis (Fig. 2B). Opposite effects of TrkA inhibitor GNF-5837 and TrkA ligand NGF on PI3K-Akt-mTOR signaling, Demonstration of a functional TrkA receptor in GOT1 Ras-Raf-MEK-ERK signaling, cell cycle components cells by NGF-mediated TrkA activation and and anti-apoptotic Bcl2 in GOT1 cells GNF-5837-mediated TrkA inhibition Recently, GNF-5837 was reported to effectively infer in Increasing concentrations of NGF caused a dose-dependent Trk-driven cancer-specific cellular pathway signaling phosphorylation of TrkA in GOT1 cells (Fig. 3A), whereas (Albaugh et al. 2012, Shargh et al. 2016). Here, we aimed different concentrations of BDNF failed to stimulate TrkA to examine which downstream signaling pathways phosphorylation (Fig. 3A), demonstrating the well-known are being affected by the Trk inhibitory function of

Table 1 IC20 and IC50 concentration values (in nM) of GNF-5837-mediated inhibition of GOT1 cell proliferation after 48, 96 and 144 h of incubation.

20% inhibitory concentration (IC20) of GNF-5837 in GOT1 cells after different times of incubation 48 h IC20 = 4.30 nM 96 h IC20 = 4.23 nM 144 h IC20 = 2.45 nM 50% inhibitory concentration (IC50) of GNF-5837 in GOT1 cells after 144 h of incubation 144 h IC50 ~ 660 nM

Cyclin D3 and cell cycle regulator p21 were downregulated (Fig. 4A and Supplementary Fig. 2). Interestingly, the expression level of the anti-apoptotic Bcl2 also decreased, suggesting an augmentation in apoptotic cell death mechanisms (Fig. 4A and Supplementary Fig. 2). In contrast to the effects observed during incubation with the pan-Trk inhibitor GNF-5837 (Fig. 4A and Supplementary Fig. 2), TrkA receptor stimulation with NGF lead to opposite effects regarding PI3K-Akt-mTOR Figure 2 signaling, Ras-Raf-MEK-ERK signaling and cell cycle (A) Base expression level of the Trk receptors (TrkA, TrkB, TrkC and regulation (Fig. 4B and Supplementary Fig. 3). Only Trkpan) in neuroendocrine tumor cells BON1, NCI-H727 and GOT1 and the positive control of human adult normal brain cells. A representative anti-apoptotic Bcl2 expression showed no major change blot out of at least 3 independently conducted Western blots is shown. in expression level upon NGF stimulation (Fig. 4B and (Only the positive control of brain tissue was analyzed less than 3 times). Supplementary Fig. 3). (B) Base expression level of the nerve growth factor (NGF) in neuroendocrine tumor cells BON1, NCI-H727 and GOT1 was evaluated by Western blot analysis. A representative blot out of three independently performed experiments is shown. (Only the positive control of brain GNF-5837 induces apoptosis and G1 cell cycle arrest tissue was analyzed less than 3 times). Cell cycle analysis confirmed the suggested GNF-5837 (5, 50 and 500 nM) after 24 h of incubation downregulation of the cell cycle and induction of (Fig. 4A and Supplementary Fig. 2). One of the most apoptosis in response to GNF-5837 treatment, by important cellular signaling pathways, which is often significantly increasing the number of cells in G1 and activated during carcinogenesis, tumor proliferation and sub-G1 phases, respectively (Fig. 7). cellular survival in NETs is the PI3K-Akt-mTOR pathway (Cantley 2002, Briest & Grabowski 2014). Here, we show Combined treatment with GNF-5837 and everolimus that key components of that pathway, such as pAkt, causes additive inhibitory effects on GOT1 cell pp70S6k, pS6 and p4EBP1 were also clearly downregulated growth mediated by PI3K-Akt-mTOR pathway and after treatment with the pan-Trk inhibitor GNF-5837 Cyclin D3 downregulation (Fig. 4A and Supplementary Fig. 2). Also another main pathway involved in NET tumorigenesis, the GOT1 cell viability was also analyzed combining the Ras-Raf-MEK-ERK pathway, was downregulated by pan-Trk inhibitor GNF-5837 with the mTOR inhibitor treatment with the pan-TRK inhibitor GNF-5837, as everolimus (Fig. 5A). The combined treatment of the expression level of pErk decreased (Fig. 4A and GNF-5837 (500 nM) with everolimus (10 nM) showed a Supplementary Fig. 2). Also the cell cycle component modest but significant enhancement over single substance

Figure 3 TrkA activation upon specific NGF stimulation and TrkA blocking upon GNF-5837 incubation in GOT1 cells. (A) A representative blot out of three of the expression lack of pTrkA after BDNF stimulation and induction of pTrkA expression by NGF stimulation in GOT1 cells is shown. GOT1 cells were treated with increasing concentrations of BDNF (0.37 nM and 3.7 nM) and NGF (0.76 nM and 7.6 nM) for 10 min. (B) Expression of pTrkA after 2 h of incubation with GNF-5837, after 10 min of incubation with NGF 7.6 nM and after combinational treatment of GNF-5837 (10 nM) and NGF (7.6 nM) in GOT1 cells. (C) Expression level of the Trkpan receptor, the TrkA receptor and its phosphorylation level after different time periods of incubation (4, 24, 48 and 72 h) with GNF-5837 (5, 50 and 500 nM). A representative blot out of three independently conducted experiments is shown.

Figure 4 (A) Effects of different concentrations of GNF-5837 (5, 50 and 500) on the expression level of different components from the cell cycle, the apoptotic response and the PI3K-Akt-mTOR and Ras-Raf-MEK-ERK pathway after 24 h of incubation in GOT1 cells. Basal protein expression level of proteins was evaluated by Western blot analysis. A representative blot out of three independently performed experiments is shown. (B) Effects of different concentrations of NGF (0.76 nM and 7.6 nM) on the expression level of different components from the cell cycle, the apoptotic response and the PI3K-Akt-mTOR and Ras-Raf-MEK-ERK pathway after 24 h of incubation in GOT1 cells. Basal protein expression level of proteins was evaluated by Western blot analysis. A representative blot out of three independently performed experiments is shown. treatment after 144 h of incubation (Fig. 5A). Other times with TrkA and pan-Trk antibodies (Fig. 6). We scaled the of incubation (72 h or 144 h) with different concentrations TMA probes for a semi-quantitative score (0 = negative, of GNF-5837 (5 nM or 500 nM) could not yield additive 1 = weak expression, 2 = moderate expression, 3 = strong effects with everolimus (data not shown). The additive expression). effect in GOT1 cell eradication was attributed to an In a total of 107 (100%) GEP-NET tumor specimens enhanced PI3K-Akt-mTOR and Cyclin D3 downregulation on the TMA, the primary tumor localization was small (Fig. 5B and Supplementary Fig. 4). Key components of intestinal (SI)-NET in 65 cases (60.7%), pancreas (p)NET the PI3K-Akt-mTOR pathway, such as pAkt, pp70S6k, in 33 (30.8%), gastric NET in 1 (0.9%), colon NET in 3 pS6 and p4EBP1, as well as the cell cycle component (2.8%), appendix NET in 3 (2.8%) and cancer of unknown Cyclin D3 showed an enhanced downregulation when primary (CUP) in 2 (1.9%) cases (Table 2). compared to single substance treatment (Fig. 5B and The antibody against TrkA stained positive (weak and Supplementary Fig. 4). The protein expression level of all moderate) exclusively in tumor samples of pancreatic other considered parameters, such as pTrkA, TrkA, pERK, primary localization as 24% (8/33) of all pNET samples p21 and Bcl2 displayed a complementary mode of action stained positive for TrkA (Table 2). A statistically significant regarding anti-proliferative signaling, when compared to relation between primary tumor localization and TrkA/ the respective single substance treatments (Fig. 5B and Trkpan expression level was proved. Supplementary Fig. 4). Using a pan-Trk antibody (detecting endogenous levels of total Trk protein encompassing TrkA, TrkB and TrkC), positive IHC expression was again found in Immunohistochemistry showed TrkA-positive all 8 previously TrkA-positive pNET samples (Table 2), staining only in pancreatic neuroendocrine as expected. In addition, the pan-Trk antibody caused tumor specimens positive staining in another 3 pNET samples which had In order to assess the prevalence of Trk receptor shown negative IHC expression in the previous staining expression in actual GEP-NET tumor samples, we with the TrkA antibody. In total, using the pan-Trk immunohistochemically stained a tissue microarray (TMA) antibody 33.3% (11/33) of all pNETs stained positive for

Figure 5 (A) Cell survival of GOT1 cells with single substance treatment of pan-Trk inhibitor GNF-5837 (500 nM) or everolimus (10 nM) and in combination (GNF-5837 + everolimus) after 144 h of incubation. The arithmetic means and s.d. of at least three independent experiments are shown. Statistical significant different results in comparison to either single substance treatment are shown, considering *P < 0.05; **P < 0.01; ***P < 0.001. (B) Effects of the combinational treatment (GNF-5837 + everolimus) on the expression level of different components from the cell cycle, the apoptotic response and the PI3K-Akt-mTOR and Ras-Raf-MEK-ERK pathway after 24 h of incubation in GOT1 cells. Basal protein expression level of proteins was evaluated by Western blot analysis. A representative blot out of three independently performed experiments is shown.

Trk family receptor expression (Table 2). In addition, 1/65 approach on major cellular signaling mechanisms such (=1.5%) SI-NETs also stained positive with the pan-Trk as the PI3K-Akt-mTOR and Ras-Raf-MEK-ERK pathway as antibody and 1/3 appendix NETs (=33.3%) (Table 2). Taken well as cell cycle components. Furthermore, we evaluated together, our results demonstrate that a total of 12.2% the prevalence of Trks immunohistochemically in a (13/107) of NET samples stained positive for receptors of large patient cohort and could assess possible predictive the Trk family (Table 2). markers for personalized molecular targeted therapy with Trk inhibitors. In the first part of this study, effects of the highly Discussion selective pan-Trk inhibitor GNF-5837 (Albaugh et al. Here, we report on the potential role of Trk inhibitors as 2012) were investigated in human neuroendocrine tumor new molecular targeting approach for NETs of the GEP cell lines in vitro. system. We evaluated the effects of the specific pan-Trk GNF-5837 significantly decreased GOT1 cell survival inhibitor GNF-5837 either alone or in a dual-targeting (Fig. 1 and Table 1) due to TrkA inhibition (Fig. 2A).

Figure 6 Expression level of TrkA in GEP-NET tumor samples. (A) TrkA strong positive (score = 3) sample. (B) TrkA negative (score = 0) sample.

Table 2 TrkA and pan-Trk expression level in different primary tumor localizations from 107 patient probes.

Primary tumor localization Small intestine Pancreas CUP Stomach Colon Appendix Total TrkA expression level TrkA negative n = 65 n = 25 n = 2 n = 1 n = 3 n = 3 n = 99 TrkA weak n = 0 n = 7 n = 0 n = 0 n = 0 n = 0 n = 7 expression TrkA moderate n = 0 n = 1 n = 0 n = 0 n = 0 n = 0 n = 1 expression Total TrkA n = 0/65 (=0%) n = 8/33 (=24.2%) n = 0/2 (=0%) n = 0/1 (=0%) n = 0/3 (=0%) n = 0/3 (=0%) n = 8/107 (=7.5%) positive pan-Trk expression level pan-Trk n = 64 n = 22 n = 2 n = 1 n = 3 n = 2 n = 94 negative pan-Trk weak n = 0 n = 9 n = 0 n = 0 n = 0 n = 0 n = 9 expression pan-Trk n = 0 n = 1 n = 0 n = 0 n = 0 n = 1 n = 2 Moderate expression pan-Trk strong n = 1 n = 1 n = 0 n = 0 n = 0 n = 0 n = 2 expression Total pan-Trk n = 1/65 n = 11/33 n = 0/2 (=0%) n = 0/1 (=0%) n = 0/3 (=0%) n = 1/3 n = 13/107 positive (=1.5%) (=33.3%) (=33.3%) (=12.2%)

Cancer of unknown primary tumor are designated ‘CUP’.

In contrast, lack of Trk expression in BON1 and due to compensatory feedback mechanisms (Fig. 3C). NCI-727 cell lines (Fig. 2A) and thus obvious independence Similar feedback loops with an upregulated re-expression of cell growth from Trk family receptors caused of TrkA after inhibition have been reported in human GNF-5837 treatment resistance in BON1 and NCI-727 cell airway smooth muscle cells (Freund-Michel & Frossard lines (Supplementary Fig. 1) and confirmed GNF-5837 2008). In Trk receptor-deficient BON1 and H727 cells, to be a highly selective pan-Trk inhibitor (Albaugh et al. NGF showed a strong expression level (Fig. 2B). Here, we 2012). The anti-proliferative effects of Trk inhibition in hypothesize that the high expression levels of NGF might neuroendocrine GOT1 tumor cells are in accordance with be due to the lack of the specific receptor (TrkA), trying to findings in other preclinical cancer cell models (Dolle et al. compensate for the missing cellular TrkA-NGF signaling 2003, Albaugh et al. 2012, Seo et al. 2013, Shargh et al. 2016). Accordingly, Trk overexpression showed to enhance tumor growth, metastasis formation and invasiveness in different tumor entities in vitro (Sortino et al. 2000, Lagadec et al. 2009). In the GOT1 cells from a concentration of 50 nM on a plateau effect is observed, where further concentration augmentation did not yield a better cell killing effect (Fig. 1). This might be explained by a maximum TrkA receptor occupation of GNF-5837 at this concentration (Maliartchouk et al. 2000). Incubation of GOT1 cells with GNF-5837 caused inhibition of TrkA activity as demonstrated by a decrease of TrkA phosphorylation status (Fig. 3A and Figure 7 C), while only specific NGF stimulation caused TrkA Cell cycle analysis measured by flow cytometry. The arithmetic means and phosphorylation (Fig. 3A and B). In contrast, GNF-5837 s.d. of at least three independent experiments are shown. Statistical significant different results in comparison to the Control DMSO are caused a significant increase of total un-phosphorylated shown, considering *P < 0.05; **P < 0.01; ***P < 0.001. A full color version TrkA receptor expression levels in GOT1 cells, possibly of this figure is available athttps://doi.org/10.1530/ERC-17-0201.

http://erc.endocrinology-journals.org © 2018 Society for Endocrinology https://doi.org/10.1530/ERC-17-0201 Published by Bioscientifica Ltd. Printed in Great Britain Downloaded from Bioscientifica.com at 09/23/2021 08:03:31PM via free access Endocrine-Related E T Aristizabal Prada, Tropomyosin receptor kinase 25:5 556 Cancer V Heinzle et al. (Trk) in NETs cascade. Tagliabue et al. found similar results when breast Supplementary Fig. 2). This fact might be explained by cancer cells with poor TrkA expression showed high NGF induction of the anti-apoptotic Bcl2 after NGF stimulation expression in order to recruit other receptors to induce and the downregulation of Bcl2 and augmentation of cancer cell growth (Tagliabue et al. 2000). the Sub-G1 population after GNF-5837 inhibition (Figs 4 The PI3K-Akt-mTOR and the Ras-Raf-MEK-ERK and 7). Studies have shown that in some systems, the pathway are often deregulated in human cancers induction of apoptotic mechanisms relate with p21 including NETs, and the importance of these pathways inhibition, counteracting its tumor-suppressive function comprises cellular oncogenic modulation mechanisms (Gartel & Tyner 2002). such as cell growth, survival, proliferation, angiogenesis Due to the PI3K-Akt-mTOR and Ras-Raf-MEK-ERK and motility (Cantley 2002, Van Gompel et al. 2005, pathway downregulation upon Trk inhibition by Kunnimalaiyaan et al. 2007, Zitzmann et al. 2007, GNF-5837 alone (Fig. 4A and Supplementary Fig. 2), 2011, Cook et al. 2010, Iida et al. 2012, Valentino et al. we aimed to further investigate the effects of 2014, Briest & Grabowski 2014). TrkA activation and GNF-5837 in a dual-targeting approach with the downstream signaling upon NGF stimulation has been mTORC1 inhibitor everolimus. The mTOR inhibitor shown in multiple cell types to mediate its proliferative everolimus has proven anti-tumoral efficacy in several and anti-apoptotic characteristics mainly through clinical phase 3 trials in patients with NET (Pavel et al. PI3K-Akt-mTOR and Ras-Raf-MEK-ERK cascade signaling 2011, Yao et al. 2016a,b). In NET cell lines, everolimus (Sofroniew et al. 2001, Huang & Reichardt 2003, has been shown to inhibit mTOR downstream signaling Rahbek et al. 2005, Jang et al. 2007, Lawrence et al. and to cause a compensatory upregulation of the Ras- 2008, Julio-Pieper et al. 2009, Amatu et al. 2016, Demir Raf-MEK-ERK signaling cascade (Zitzmann et al. 2010, et al. 2016, Pediaditakis et al. 2016). In accordance to Wagle et al. 2014, Capozzi et al. 2015, He et al. 2016, these data, in GOT1 cells, NGF stimulation induced the Vandamme et al. 2016, Zatelli et al. 2016). Furthermore, PI3K-Akt-mTOR and Ras-Raf-MEK-ERK signaling cascade development of resistance against everolimus during (Fig. 4B), while inhibition of TrkA signaling by the long-term treatment is often observed in NETs and pan-Trk inhibitor GNF-5837 caused downregulation possibly caused by feedback loops on the PI3K-Akt axis of both downstream signaling pathways (Fig. 4A). (Zitzmann et al. 2010, Wagle et al. 2014, Capozzi et al. Accordingly, in a colorectal cancer cell line (KM12) and 2015, He et al. 2016, Vandamme et al. 2016, Zatelli et al. a LCNEC cell line, pan-Trk inhibition with AZD7451 2016). Thus, the potential benefit of a dual-targeting leads to a cellular survival decrease through pErk or pAkt approach is not only to enhance single substance downregulation (Tatematsu et al. 2014). As a consequence to treatment efficacy, but also to prevent and/or counteract PI3K-Akt-mTOR and Ras-Raf-MEK-ERK alteration, resistance mechanisms. We found that the combinational downstream cell cycle component Cyclin D3 was approach of the Trk inhibitor GNF-5837 and the mTOR upregulated upon NGF stimulation or downregulated inhibitor everolimus caused additive anti-proliferative upon GNF-5837 inhibition, respectively (Fig. 4). effects (Fig. 5) and agonistically downregulated the Furthermore, a significant population increase in the cell PI3K-Akt-mTOR pathway, as well as Cyclin D3 (Fig. 5B and cycle phase G1 confirms a cell cycle downregulation upon Supplementary Fig. 4). In a Trk-induced leukemia mouse- GNF-5837 treatment (Fig. 7). GOT1 cells showed not only model, the tumor cells showed a high susceptibility a strong TrkA expression but also a slight expression of toward the mTORC1 inhibitor rapamycin and RAD001 TrkC (Fig. 2A). Besides inhibition of TrkA, also inhibition (everolimus) via PI3K-Akt-mTOR pathway and of TrkC, might contribute to Ras-Raf-MEK-ERK and PI3K- Ras-Raf-MEK-ERK downregulation (Rhein et al. 2011). Akt-mTOR pathway inhibition following incubation In the second part of this study, we examined the with the pan-Trk inhibitor GNF-5837. In adenoid cystic expression of Trk family receptors in a tissue microarray carcinoma cells, TrkC signaling was activated and (TMA) encompassing 107 tumor probes of patients with required its ligand NT-3 to stimulate invasive behavior GEP-NETs. through Ras-Raf-MEK-ERK and PI3K-Akt-mTOR signaling Using a specific TrkA antibody, a positive staining (Ivanov et al. 2013). in 8/107 tumor probes was exclusively due to the sub- Interestingly, the cyclin-dependent kinase inhibitor group of pNETs with a positive staining in 8/33 (24%) of p21 is upregulated upon NGF stimulation (Fig. 4B all pNETs (Table 2). Using a pan-Trk antibody (detecting and Supplementary Fig. 3) and downregulated upon endogenous levels of total Trk protein encompassing TrkA, Trk inhibition by GNF-5837 treatment (Fig. 4A and TrkB and TrkC) a positive staining in 13/107 tumor probes

http://erc.endocrinology-journals.org © 2018 Society for Endocrinology https://doi.org/10.1530/ERC-17-0201 Published by Bioscientifica Ltd. Printed in Great Britain Downloaded from Bioscientifica.com at 09/23/2021 08:03:31PM via free access Endocrine-Related E T Aristizabal Prada, Tropomyosin receptor kinase 25:5 557 Cancer V Heinzle et al. (Trk) in NETs was mainly due to the subset of pNETs with a positive Declaration of interest staining in 11/33 (33%) of all pNETs (Table 2). These data C J A has received research contracts (Ipsen, Novartis), lecture honorarium demonstrate a high prevalence of Trk receptor expression (Ipsen, Novartis, Pfizer, Falk) and advisory board honorarium (Novartis). in pancreatic NETs. Therefore, immunohistochemical or The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work. molecular screening for Trk expression in pNETs might be useful as a predictive marker to select patients with pNETs expressing Trks for molecular targeted therapy with Trk Funding inhibitors. Whether Trk overexpression or overactivation This study has been funded by the Theranostic Research Network, Bad caused by chromosomal rearrangements and NTRK Berka, Germany – Wilhelm und Ingeburg Dinse Gedächtnis-Stiftung. The gene fusions (Amatu et al. 2016) is also existent in NET biosampling of this study has been initially funded within NeoExNET (Network of Excellence for Neuroendocrine Tumors in Munich). NeoExNET should be investigated in further studies. Interestingly, is a national database for the evaluation of diagnostics, treatment and in the malignant neuroendocrine Merkel cell carcinoma, outcome in neuroendocrine tumors. NeoExNET is supported by the expression of TrkA was found in all tested tumor specimens German Federal Ministry of Education and Research (BMBF; m4 Cluster – Personalized Medicine and Targeted Therapies Leading-Edge Cluster (36/36), underlining a possible therapeutic relevance Munich, 16EX1221J, 01EX1020E) and by an unrestricted educational grant of Trk blockage in NETs not only from the GEP system of Novartis Pharma GmbH, Nürnberg, Germany. Members of the NeoExNET- (Wehkamp et al. 2017). Study group include: Principal investigator: Günter K Stalla (Max Planck Institute of Psychiatry, Munich). Steering Committee: Felix Beuschlein A limitation of our study is the fact that currently (Ludwig Maximilian University, Munich), Christoph Auernhammer (Ludwig available human neuroendocrine tumor cell lines might Maximilian University, Munich), Klaus A Kuhn (Technische Universität not reflect the actual biological expression state of Trk München, Munich). receptors in neuroendocrine tumors in vivo. While the SI carcinoid cell line GOT1 was the only tested cell line References expressing a functional Trk receptor (Figs 1, 2A and 3), Albaugh P, Fan Y, Mi Y, Sun F, Adrian F, Li N, Jia Y, Sarkisova Y, the TMA analysis revealed Trk expression in SI NETs Kreusch A, Hood T, et al. 2012 Discovery of GNF-5837, a selective to be a very rare event as only 1/65 (1.5%) of SI NET TRK inhibitor with efficacy in rodent cancer tumor models. ACS tumor probes showed a positive staining with the pan- Medicinal Chemistry Letters 3 140–145. (https://doi.org/10.1021/ ml200261d) Trk antibody (Table 2). Further investigation regarding Amatu A, Sartore-Bianchi A & Siena S 2016 NTRK gene fusions as novel the effects of GNF-5837 in NETs is required in order targets of cancer therapy across multiple tumour types. ESMO Open 1 to validate molecular target credentials and clinical e000023. (https://doi.org/10.1136/esmoopen-2015-000023) Ardini E, Menichincheri M, Banfi P, Bosotti R, De Ponti C, Pulci R, relevance. Ballinari D, Ciomei M, Texido G, Degrassi A, et al. 2016 Entrectinib, Taken together our findings, the highly selective a Pan-TRK, ROS1, and ALK inhibitor with activity in multiple pan-Trk inhibitor GNF-5837 presents significant anti- molecularly defined cancer indications. Molecular Cancer Therapeutics 15 628–639. (https://doi.org/10.1158/1535-7163.MCT-15-0758) proliferative efficacy in human neuroendocrine GOT1 Auernhammer CJ & Goke B 2011 Therapeutic strategies for advanced cells in vitro. Inhibition of TrkA signaling in GOT1 cells neuroendocrine carcinomas of jejunum/ileum and pancreatic origin. caused subsequent downregulation of the PI3K-Akt- Gut 60 1009–1021. (https://doi.org/10.1136/gut.2009.204453) Barbacid M 1994 The Trk family of neurotrophin receptors. Journal of mTOR and Ras-Raf-MEK-ERK signaling cascades, as Neurobiology 25 1386–1403. (https://doi.org/10.1002/neu.480251107) well as the cell cycle and upregulation of apoptosis. A Bibel M & Barde YA 2000 Neurotrophins: key regulators of cell fate and dual-targeting approach using the pan-Trk inhibitor cell shape in the vertebrate nervous system. Genes and Development 14 2919–2937. (https://doi.org/10.1101/gad.841400) GNF-5837 plus the mTOR inhibitor everolimus showed Bodei L, Cremonesi M, Kidd M, Grana CM, Severi S, Modlin IM & additive inhibitory effects on PI3K-Akt-mTOR pathway Paganelli G 2014 Peptide receptor radionuclide therapy for advanced and cyclin D3. Tissue microarray analysis revealed Trk neuroendocrine tumors. Thoracic Surgery Clinics 24 333–349. (https:// doi.org/10.1016/j.thorsurg.2014.04.005) expression in 24% (TrkA)/33% (pan-Trk) of pancreatic Briest F & Grabowski P 2014 PI3K-AKT-mTOR-signaling and beyond: the NETs. Immunohistochemical or molecular screening for complex network in gastroenteropancreatic neuroendocrine Trk expression in pNETs might be useful as a predictive neoplasms. Theranostics 4 336–365. (https://doi.org/10.7150/ thno.7851) marker to select patients with pNETs expressing Trks for Cakir M & Grossman A 2011 The molecular pathogenesis and molecular targeted therapy with Trk inhibitors. management of bronchial carcinoids. Expert Opinion on Therapeutic Targets 15 457–491. (https://doi.org/10.1517/14728222.2011.555403) Cantley LC 2002 The phosphoinositide 3-kinase pathway. Science 296 1655–1657. (https://doi.org/10.1126/science.296.5573.1655) Supplementary data Capozzi M, Caterina I, De Divitiis C, von Arx C, Maiolino P, Tatangelo F, This is linked to the online version of the paper at https://doi.org/10.1530/ Cavalcanti E, Di Girolamo E, Iaffaioli RV, Scala S, et al. 2015 ERC-17-0201. Everolimus and pancreatic neuroendocrine tumors (PNETs): activity,

resistance and how to overcome it. International Journal of Surgery 21 robust neurotrophic activity, prevents neuronal cell death. PNAS 104 (Supplement 1) S89–S94. (https://doi.org/10.1016/j.ijsu.2015.06.064) 16329–16334. (https://doi.org/10.1073/pnas.0706662104) Cives M & Strosberg J 2017 Radionuclide therapy for neuroendocrine Julio-Pieper M, Lozada P, Tapia V, Vega M, Miranda C, Vantman D, tumors. Current Oncology Reports 19 9. (https://doi.org/10.1007/ Ojeda SR & Romero C 2009 Nerve growth factor induces vascular s11912-017-0567-8) endothelial growth factor expression in granulosa cells via a trkA Clary DO & Reichardt LF 1994 An alternatively spliced form of the receptor/mitogen-activated protein kinase-extracellularly regulated nerve growth factor receptor TrkA confers an enhanced response to kinase 2-dependent pathway. Journal of Clinical Endocrinology and neurotrophin 3. PNAS 91 11133–11137. (https://doi.org/10.1073/ Metabolism 94 3065–3071. (https://doi.org/10.1210/jc.2009-0542) pnas.91.23.11133) Kaplan DR, Hempstead BL, Martin-Zanca D, Chao MV & Parada LF 1991 Cook MR, Pinchot SN, Jaskula-Sztul R, Luo J, Kunnimalaiyaan M & The trk proto-oncogene product: a signal transducing receptor for Chen H 2010 Identification of a novel Raf-1 pathway activator that nerve growth factor. Science 252 554–558. (https://doi.org/10.1126/ inhibits gastrointestinal carcinoid cell growth. Molecular Cancer science.1850549) Therapeutics 9 429–437. (https://doi.org/10.1158/1535-7163.MCT-09- Karpathakis A, Dibra H, Pipinikas C, Feber A, Morris T, Francis J, 0718) Oukrif D, Mandair D, Pericleous M, Mohmaduvesh M, et al. 2017 Davidson B, Lazarovici P, Ezersky A, Nesland JM, Berner A, Risberg B, Progressive epigenetic dysregulation in neuroendocrine tumour liver Trope CG, Kristensen GB, Goscinski M, van de Putte G, et al. 2001 metastases. Endocrine-Related Cancer 24 L21–L25. (https://doi. Expression levels of the nerve growth factor receptors TrkA and p75 org/10.1530/ERC-16-0419) in effusions and solid tumors of serous ovarian carcinoma patients. Khotskaya YB, Holla VR, Farago AF, Mills Shaw KR, Meric-Bernstam F & Clinical Cancer Research 7 3457–3464. Hong DS 2017 Targeting TRK family proteins in cancer. Pharmacology Demir IE, Tieftrunk E, Schorn S, Friess H & Ceyhan GO 2016 Nerve and Therapeutics 173 58–66. (https://doi.org/10.1016/j. growth factor & TrkA as novel therapeutic targets in cancer. pharmthera.2017.02.006) Biochimica et Biophysica Acta 1866 37–50. (https://doi.org/10.1016/j. Kidd M, Modlin I & Oberg K 2016 Towards a new classification of bbcan.2016.05.003) gastroenteropancreatic neuroendocrine neoplasms. Nature Reviews Dolle L, El Yazidi-Belkoura I, Adriaenssens E, Nurcombe V & Clinical Oncology 13 691–705. (https://doi.org/10.1038/ Hondermarck H 2003 Nerve growth factor overexpression and nrclinonc.2016.85) autocrine loop in breast cancer cells. Oncogene 22 5592–5601. Klein R, Jing SQ, Nanduri V, O’Rourke E & Barbacid M 1991 The trk (https://doi.org/10.1038/sj.onc.1206805) proto-oncogene encodes a receptor for nerve growth factor. Cell 65 Drilon A, Siena S, Ou SI, Patel M, Ahn MJ, Lee J, Bauer TM, Farago AF, 189–197. (https://doi.org/10.1016/0092-8674(91)90419-Y) Wheler JJ, Liu SV, et al. 2017 Safety and antitumor activity of the Knosel T, Emde A, Schluns K, Chen Y, Jurchott K, Krause M, Dietel M & multi-targeted Pan-TRK, ROS1, and ALK Inhibitor Entrectinib Petersen I 2005 Immunoprofiles of 11 biomarkers using tissue (RXDX-101): combined results from two phase 1 trials (ALKA-372- microarrays identify prognostic subgroups in colorectal cancer. 001 and STARTRK-1). Cancer Discovery 7 400–409. (https://doi. Neoplasia 7 741–747. (https://doi.org/10.1593/neo.05178) org/10.1158/2159-8290.CD-16-1237) Knosel T, Emde V, Schluns K, Schlag PM, Dietel M & Petersen I 2006 Evers BM, Townsend CM Jr, Upp JR, Allen E, Hurlbut SC, Kim SW, Cytokeratin profiles identify diagnostic signatures in colorectal Rajaraman S, Singh P, Reubi JC & Thompson JC 1991 Establishment cancer using multiplex analysis of tissue microarrays. Cellular and characterization of a human carcinoid in nude mice and effect Oncology 28 167–175. (https://doi.org/10.1155/2006/354295) of various agents on tumor growth. Gastroenterology 101 303–311. Kolby L, Bernhardt P, Ahlman H, Wangberg B, Johanson V, Wigander A, (https://doi.org/10.1016/0016-5085(91)90004-5) Forssell-Aronsson E, Karlsson S, Ahren B, Stenman G, et al. 2001 A Freund-Michel V & Frossard N 2008 Overexpression of functional TrkA transplantable human carcinoid as model for somatostatin receptor- receptors after internalisation in human airway smooth muscle cells. mediated and amine transporter-mediated radionuclide uptake. Biochimica et Biophysica Acta 1783 1964–1971. (https://doi. American Journal of Pathology 158 745–755. (https://doi.org/10.1016/ org/10.1016/j.bbamcr.2008.05.014) S0002-9440(10)64017-5) Gartel AL & Tyner AL 2002 The role of the cyclin-dependent kinase Kunnimalaiyaan M, Ndiaye M & Chen H 2007 Neuroendocrine tumor inhibitor p21 in apoptosis. Molecular Cancer Therapeutics 1 639–649. cell growth inhibition by ZM336372 through alterations in multiple Grozinsky-Glasberg S, Shimon I & Rubinfeld H 2012 The role of cell signaling pathways. Surgery 142 959–964. (https://doi.org/10.1016/j. lines in the study of neuroendocrine tumors. Neuroendocrinology 96 surg.2007.09.020) 173–187. (https://doi.org/10.1159/000338793) Lagadec C, Meignan S, Adriaenssens E, Foveau B, Vanhecke E, Romon R, He K, Chen D, Ruan H, Li X, Tong J, Xu X, Zhang L & Yu J 2016 Toillon RA, Oxombre B, Hondermarck H & Le Bourhis X 2009 TrkA BRAFV600E-dependent Mcl-1 stabilization leads to everolimus overexpression enhances growth and metastasis of breast cancer resistance in colon cancer cells. Oncotarget 7 47699–47710. (https:// cells. Oncogene 28 1960–1970. (https://doi.org/10.1038/onc.2009.61) doi.org/10.18632/oncotarget.10277) Lawrence MC, McGlynn K, Shao C, Duan L, Naziruddin B, Levy MF & Huang EJ & Reichardt LF 2003 Trk receptors: roles in neuronal signal Cobb MH 2008 Chromatin-bound mitogen-activated protein kinases transduction. Annual Review of Biochemistry 72 609–642. (https://doi. transmit dynamic signals in transcription complexes in beta-cells. org/10.1146/annurev.biochem.72.121801.161629) PNAS 105 13315–13320. (https://doi.org/10.1073/pnas.0806465105) Iida S, Miki Y, Ono K, Akahira J, Nakamura Y, Suzuki T & Sasano H 2012 Maliartchouk S, Debeir T, Beglova N, Cuello AC, Gehring K & Synergistic anti-tumor effects of RAD001 with MEK inhibitors in Saragovi HU 2000 Genuine monovalent ligands of TrkA nerve neuroendocrine tumors: a potential mechanism of therapeutic growth factor receptors reveal a novel pharmacological mechanism limitation of mTOR inhibitor. Molecular and Cellular Endocrinology of action. Journal of Biological Chemistry 275 9946–9956. (https://doi. 350 99–106. (https://doi.org/10.1016/j.mce.2011.11.024) org/10.1074/jbc.275.14.9946) Ivanov SV, Panaccione A, Brown B, Guo Y, Moskaluk CA, Wick MJ, Martin-Zanca D, Oskam R, Mitra G, Copeland T & Barbacid M 1989 Brown JL, Ivanova AV, Issaeva N, El-Naggar AK, et al. 2013 TrkC Molecular and biochemical characterization of the human trk proto- signaling is activated in adenoid cystic carcinoma and requires NT-3 oncogene. Molecular and Cellular Biology 9 24–33. (https://doi. to stimulate invasive behavior. Oncogene 32 3698–3710. (https://doi. org/10.1128/MCB.9.1.24) org/10.1038/onc.2012.377) McGregor LM, McCune BK, Graff JR, McDowell PR, Romans KE, Jang SW, Okada M, Sayeed I, Xiao G, Stein D, Jin P & Ye K 2007 Yancopoulos GD, Ball DW, Baylin SB & Nelkin BD 1999 Roles of trk Gambogic amide, a selective agonist for TrkA receptor that possesses family neurotrophin receptors in medullary thyroid carcinoma

development and progression. PNAS 96 4540–4545. (https://doi. neuroendocrine tumors. Langenbeck’s Archives of Surgery 396 org/10.1073/pnas.96.8.4540) 273–298. (https://doi.org/10.1007/s00423-011-0739-1) Modlin IM, Moss SF, Chung DC, Jensen RT & Snyderwine E 2008a Seo JH, Jung KH, Son MK, Yan HH, Ryu YL, Kim J, Lee JK, Hong S & Priorities for improving the management of gastroenteropancreatic Hong SS 2013 Anti-cancer effect of HS-345, a new tropomyosin- neuroendocrine tumors. Journal of the National Cancer Institute 100 related kinase A inhibitor, on human pancreatic cancer. Cancer 1282–1289. (https://doi.org/10.1093/jnci/djn275) Letters 338 271–281. (https://doi.org/10.1016/j.canlet.2013.04.002) Modlin IM, Oberg K, Chung DC, Jensen RT, de Herder WW, Thakker RV, Shargh VH, Hondermarck H & Liang M 2016 Albumin hybrid Caplin M, Delle Fave G, Kaltsas GA, Krenning EP, et al. 2008b nanoparticles loaded with tyrosine kinase A inhibitor GNF-5837 for Gastroenteropancreatic neuroendocrine tumours. Lancet Oncology 9 targeted inhibition of breast cancer cell growth and invasion. 61–72. (https://doi.org/10.1016/S1470-2045(07)70410-2) International Journal of Pharmaceutics 515 527–534. (https://doi. Numakawa T, Suzuki S, Kumamaru E, Adachi N, Richards M & Kunugi H org/10.1016/j.ijpharm.2016.10.057) 2010 BDNF function and intracellular signaling in neurons. Histology Sofroniew MV, Howe CL & Mobley WC 2001 Nerve growth factor and Histopathology 25 237–258. (https://doi.org/10.14670/HH-25.237) signaling, neuroprotection, and neural repair. Annual Review of Odate S, Nakamura K, Onishi H, Kojima M, Uchiyama A, Nakano K, Neuroscience 24 1217–1281. (https://doi.org/10.1146/annurev. Kato M, Tanaka M & Katano M 2013 TrkB/BDNF signaling pathway neuro.24.1.1217) is a potential therapeutic target for pulmonary large cell Sortino MA, Condorelli F, Vancheri C, Chiarenza A, Bernardini R, neuroendocrine carcinoma. Lung Cancer 79 205–214. (https://doi. Consoli U & Canonico PL 2000 Mitogenic effect of nerve growth org/10.1016/j.lungcan.2012.12.004) factor (NGF) in LNCaP prostate adenocarcinoma cells: role of the Oelmann E, Sreter L, Schuller I, Serve H, Koenigsmann M, high- and low-affinity NGF receptors. Molecular Endocrinology 14 Wiedenmann B, Oberberg D, Reufi B, Thiel E & Berdel WE 1995 124–136. (https://doi.org/10.1210/mend.14.1.0402) Nerve growth factor stimulates clonal growth of human lung cancer Spampatti M, Vlotides G, Spottl G, Maurer J, Goke B & cell lines and a human glioblastoma cell line expressing high-affinity Auernhammer CJ 2014 Aspirin inhibits cell viability and mTOR nerve growth factor binding sites involving tyrosine kinase signaling. downstream signaling in gastroenteropancreatic and Cancer Research 55 2212–2219. bronchopulmonary neuroendocrine tumor cells. World Journal of Osborne JK, Larsen JE, Shields MD, Gonzales JX, Shames DS, Sato M, Gastroenterology 20 10038–10049. (https://doi.org/10.3748/wjg.v20. Kulkarni A, Wistuba II, Girard L, Minna JD, et al. 2013 NeuroD1 i29.10038) regulates survival and migration of neuroendocrine lung carcinomas Tagliabue E, Castiglioni F, Ghirelli C, Modugno M, Asnaghi L, via signaling molecules TrkB and NCAM. PNAS 110 6524–6529. Somenzi G, Melani C & Menard S 2000 Nerve growth factor (https://doi.org/10.1073/pnas.1303932110) cooperates with p185(HER2) in activating growth of human breast Passiglia F, Caparica R, Giovannetti E, Giallombardo M, Listi A, Diana P, carcinoma cells. Journal of Biological Chemistry 275 5388–5394. Cirrincione G, Caglevic C, Raez LE, Russo A, et al. 2016 The (https://doi.org/10.1074/jbc.275.8.5388) potential of neurotrophic tyrosine kinase (NTRK) inhibitors for Tatematsu T, Sasaki H, Shimizu S, Okuda K, Shitara M, Hikosaka Y, treating lung cancer. Expert Opinion on Investigational Drugs 25 Moriyama S, Yano M, Brown J & Fujii Y 2014 Investigation of 385–392. (https://doi.org/10.1517/13543784.2016.1152261) neurotrophic tyrosine kinase receptor 1 fusions and neurotrophic Pavel ME & Sers C 2016 WOMEN IN CANCER THEMATIC REVIEW: tyrosine kinase receptor family expression in non-small-cell lung systemic therapies in neuroendocrine tumors and novel approaches cancer and sensitivity to AZD7451 in vitro. Molecular and Clinical toward personalized medicine. Endocrine-Related Cancer 23 Oncology 2 725–730. (https://doi.org/10.3892/mco.2014.318) T135–T154. (https://doi.org/10.1530/ERC-16-0370) Valentino JD, Li J, Zaytseva YY, Mustain WC, Elliott VA, Kim JT, Pavel ME, Hainsworth JD, Baudin E, Peeters M, Horsch D, Winkler RE, Harris JW, Campbell K, Weiss H, Wang C, et al. 2014 Cotargeting the Klimovsky J, Lebwohl D, Jehl V, Wolin EM, et al. 2011 Everolimus PI3K and RAS pathways for the treatment of neuroendocrine tumors. plus octreotide long-acting repeatable for the treatment of advanced Clinical Cancer Research 20 1212–1222. (https://doi. neuroendocrine tumours associated with carcinoid syndrome org/10.1158/1078-0432.CCR-13-1897) (RADIANT-2): a randomised, placebo-controlled, phase 3 study. Lancet Van Gompel JJ, Kunnimalaiyaan M, Holen K & Chen H 2005 ZM336372, 378 2005–2012. (https://doi.org/10.1016/S0140-6736(11)61742-X) a Raf-1 activator, suppresses growth and neuroendocrine hormone Pediaditakis I, Efstathopoulos P, Prousis KC, Zervou M, Arevalo JC, levels in carcinoid tumor cells. Molecular Cancer Therapeutics 4 Alexaki VI, Nikoletopoulou V, Karagianni E, Potamitis C, 910–917. (https://doi.org/10.1158/1535-7163.MCT-04-0334) Tavernarakis N, et al. 2016 Selective and differential interactions of van Vliet EI, Teunissen JJ, Kam BL, de Jong M, Krenning EP & BNN27, a novel C17-spiroepoxy steroid derivative, with TrkA Kwekkeboom DJ 2013 Treatment of gastroenteropancreatic receptors, regulating neuronal survival and differentiation. neuroendocrine tumors with peptide receptor radionuclide Neuropharmacology 111 266–282. (https://doi.org/10.1016/j. therapy. Neuroendocrinology 97 74–85. (https://doi. neuropharm.2016.09.007) org/10.1159/000335018) Rahbek UL, Dissing S, Thomassen C, Hansen AJ & Tritsaris K 2005 Nerve Vandamme T, Beyens M, de Beeck KO, Dogan F, van Koetsveld PM, growth factor activates aorta endothelial cells causing PI3K/Akt- and Pauwels P, Mortier G, Vangestel C, de Herder W, Van Camp G, et al. ERK-dependent migration. Pflügers Archiv 450 355–361. (https://doi. 2016 Long-term acquired everolimus resistance in pancreatic org/10.1007/s00424-005-1436-0) neuroendocrine tumours can be overcome with novel PI3K-AKT- Raymond E, Dahan L, Raoul JL, Bang YJ, Borbath I, Lombard-Bohas C, mTOR inhibitors. British Journal of Cancer 114 650–658. (https://doi. Valle J, Metrakos P, Smith D, Vinik A, et al. 2011 Sunitinib malate for org/10.1038/bjc.2016.25) the treatment of pancreatic neuroendocrine tumors. New England Wagle N, Grabiner BC, Van Allen EM, Amin-Mansour A, Taylor- Journal of Medicine 364 501–513. (https://doi.org/10.1056/ Weiner A, Rosenberg M, Gray N, Barletta JA, Guo Y, Swanson SJ, et NEJMoa1003825) al. 2014 Response and acquired resistance to everolimus in Rhein M, Schwarzer A, Yang M, Kaever V, Brugman M, Meyer J, Ganser A, anaplastic thyroid cancer. New England Journal of Medicine 371 Baum C & Li Z 2011 Leukemias induced by altered TRK-signaling are 1426–1433. (https://doi.org/10.1056/NEJMoa1403352) sensitive to mTOR inhibitors in preclinical models. Annals of Wehkamp U, Stern S, Kruger S, Hauschild A, Rocken C & Egberts F 2017 Hematology 90 283–292. (https://doi.org/10.1007/s00277-010-1065-3) Tropomyosin receptor kinase A expression on merkel cell carcinoma Schimmack S, Svejda B, Lawrence B, Kidd M & Modlin IM 2011 The cells. JAMA Dermatology 153 1166–1169. (https://doi.org/10.1001/ diversity and commonalities of gastroenteropancreatic jamadermatol.2017.2495)

Yao JC, Hassan M, Phan A, Dagohoy C, Leary C, Mares JE, Abdalla EK, Zatelli MC, Fanciulli G, Malandrino P, Ramundo V, Faggiano A & Fleming JB, Vauthey JN, Rashid A, et al. 2008 One hundred years Colao A 2016 Predictive factors of response to mTOR inhibitors in after ‘carcinoid’: epidemiology of and prognostic factors for neuroendocrine tumours. Endocrine-Related Cancer 23 R173–R183. neuroendocrine tumors in 35,825 cases in the United States. Journal (https://doi.org/10.1530/ERC-15-0413) of Clinical Oncology 26 3063–3072. (https://doi.org/10.1200/ Zhu ZW, Friess H, Wang L, Di Mola FF, Zimmermann A & Buchler MW JCO.2007.15.4377) 2000 Down-regulation of nerve growth factor in poorly Yao JC, Shah MH, Ito T, Bohas CL, Wolin EM, Van Cutsem E, differentiated and advanced human esophageal cancer. Anticancer Hobday TJ, Okusaka T, Capdevila J, de Vries EG, et al. 2011 Research 20 125–132. Everolimus for advanced pancreatic neuroendocrine tumors. New Zitzmann K, De Toni EN, Brand S, Goke B, Meinecke J, Spottl G, England Journal of Medicine 364 514–523. (https://doi.org/10.1056/ Meyer HH & Auernhammer CJ 2007 The novel mTOR inhibitor NEJMoa1009290) RAD001 (everolimus) induces antiproliferative effects in human Yao JC, Fazio N, Singh S, Buzzoni R, Carnaghi C, Wolin E, Tomasek J, pancreatic neuroendocrine tumor cells. Neuroendocrinology 85 54–60. Raderer M, Lahner H, Voi M, et al. 2016a Everolimus for the (https://doi.org/10.1159/000100057) treatment of advanced, non-functional neuroendocrine tumours of Zitzmann K, Ruden J, Brand S, Goke B, Lichtl J, Spottl G & the lung or gastrointestinal tract (RADIANT-4): a randomised, Auernhammer CJ 2010 Compensatory activation of Akt in response placebo-controlled, phase 3 study. Lancet 387 968–977. (https://doi. to mTOR and Raf inhibitors – a rationale for dual-targeted therapy org/10.1016/S0140-6736(15)00817-X) approaches in neuroendocrine tumor disease. Cancer Letters 295 Yao JC, Pavel M, Lombard-Bohas C, Van Cutsem E, Voi M, Brandt U, 100–109. (https://doi.org/10.1016/j.canlet.2010.02.018) He W, Chen D, Capdevila J, de Vries EG, et al. 2016b Everolimus for Zitzmann K, de Toni E, von Ruden J, Brand S, Goke B, Laubender RP & the treatment of advanced pancreatic neuroendocrine tumors: Auernhammer CJ 2011 The novel Raf inhibitor Raf265 decreases overall survival and circulating biomarkers from the randomized, Bcl-2 levels and confers TRAIL-sensitivity to neuroendocrine tumour phase III RADIANT-3 Study. Journal of Clinical Oncology 34 cells. Endocrine-Related Cancer 18 277–285. (https://doi.org/10.1530/ 3609–3913. (https://doi.org/10.1200/JCO.2016.68.0702) ERC-10-0108)

Received in final form 9 March 2018 Accepted 21 March 2018 Accepted Preprint published online 21 March 2018